10 weeks to feature freeze (Pending Work)

jd@commandprompt.com

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: Updateable cursors

FAST PostgreSQL wrote:

We are trying to develop the updateable cursors functionality into
Postgresql. I have given below details of the design and also issues we are
facing. Looking forward to the advice on how to proceed with these issues.

Rgds,
Arul Shaji

Would this be something that you would hope to submit for 8.3?

Joshua D. Drake

1. Introduction
--------------
This is a combined proposal and design document for adding updatable
(insensitive) cursor capability to the PostgreSQL database.
There have already been a couple of previous proposals since 2003 for
implementing this feature so there appears to be community interest in doing
so. This will enable the following constructs to be processed:

UPDATE <table_name> SET value_list WHERE CURRENT OF <cursor_name>
DELETE FROM <table_name> WHERE CURRENT OF <cursor_name>

This has the effect of users being able to update or delete specific rows of
a table, as defined by the row currently fetched into the cursor.

2. Overall Conceptual Design
-----------------------------
The design is considered from the viewpoint of progression of a command
through the various stages of processing, from changes to the file ï¿½gram.yï¿½
to implement the actual grammar changes, through to changes in the Executor
portion of the database architecture.

2.1 Changes to the Grammar
------------------------------
The following changes will be done to the PostgreSQL grammar:

UPDATE statement has the option ï¿½WHERE CURRENT OF <cursor_name>ï¿½ added
DELETE statement has the option ï¿½WHERE CURRENT OF <cursor_name>ï¿½ added

The cursor_name data is held in the UpdateStmt and DeleteStmt structures and
contains just the name of the cursor.

The pl/pgsql grammar changes in the same manner.

The word CURRENT will be added to the ScanKeywords array in keywords.c.

2.2 Changes to Affected Data Structures
------------------------------------------
The following data structures are affected by this change:

Portal structure, QueryDesc structure, the UpdateStmt and DeleteStmt
structures

The Portal will contain a list of structures of relation ids and tuple ids
relating to the tuple held in the QueryDesc structure. There will be one
entry in the relation and tuple id list for each entry in the relation-list
of the statement below:

DECLARE <cursor_name> [WITH HOLD] SELECT FOR UPDATE OF <relation-list>

The QueryDesc structure will contain the relation id and the tuple id
relating to the tuple obtained via the FETCH command so that it can be
propagated back to the Portal for storage in the list described above.

The UpdateStmt and DeleteStmt structures have the cursor name added so that
the information is available for use in obtaining the portal structure
related to the cursor previously opened via the DECLARE CURSOR request.

2.3 Changes to the SQL Parser
------------------------------------
At present, although the FOR UPDATE clause of the DECLARE CURSOR command has
been present in the grammar, it causes an error message later in the
processing since cursors are currently not updatable. This now needs to
change. The ï¿½FOR UPDATEï¿½ clause has to be valid, but not the ï¿½FOR SHAREï¿½
clause.

The relation names that follow the ï¿½FOR UPDATEï¿½ clause will be added to the
rtable in the Query structure and identified by means of the rowMarks array.
In the case of an updatable cursor the FOR SHARE option is not allowed
therefore all entries in the rtable that are identified by the rowMarks array
must relate to tables that are FOR UPDATE.

In the UPDATE or DELETE statements the ï¿½WHERE CURRENT OF <cursor_name>ï¿½
clause results in the cursor name being placed in the UpdateStmt or
DeleteStmt structure. During the processing of the functions -
transformDeleteStmt() and transformUpdateStmt() - the cursor name is used to
obtain a pointer to the related Portal structure and the tuple affected by
the current UPDATE or DELETE statement is extracted from the Portal, where it
has been placed as the result of a previous FETCH request. At this point all
the information for the UPDATE or DELETE statement is available so the
statements can be transformed into standard UPDATE or DELETE statements and
sent for re-write/planning/execution as usual.

2.4 Changes to the Optimizer
------------------------------
There is a need to add a TidScan node to planning UPDATE / DELETE statements
where the statements are ï¿½UPDATE / DELETE at positionï¿½. This is to enable the
tuple ids of the tuples in the tables relating to the query to be obtained.
There will need to be a new mechanism to achieve this, as at present, a Tid
scan is done only if there is a standard WHERE condition on update or delete
statements to provide Tid qualifier data.

2.5 Changes to the Executor
-------------------------------
There are various options that have been considered for this part of the
enhancement. These are described in the sections below.

We would like to hear opinions on which option is the best way to go or if
none of these is acceptable, any alternate ideas ?

Option 1 MVCC Via Continuous Searching of Database

The Executor is to be changed in the following ways:
1) When the FETCH statement is executed the id of the resulting tuple is
extracted and passed back to the Portal structure to be saved to indicate the
cursor is currently positioned on a tuple.
2) When the UPDATE or DELETE request is executed the tuple id previously
FETCHed is held in the QueryDesc structure so that it can be compared with
the tuple ids returned from the TidScan node processed prior to the actual
UPDATE / DELETE node in the plan. This enables a decision to be made as to
whether the tuple held in the cursor is visible to the UPDATE / DELETE
request according to the rules of concurrency. The result is that, at the
cost of repeatedly searching the database at each UPDATE / DELETE command,
the hash table is no longer required.
This approach has the advantage that there is no hash table held in memory or
on disk so it will not be memory intensive but will be processing intensive.

This is a good ï¿½one-offï¿½ solution to the problem and, taken in isolation is
probably the best approach. However, if one considers the method(s) used in
other areas of PostgreSQL, it is probably not the best solution. This option
will probably not be used further.

Option 2 MVCC via New Snapshot

The executor can be changed by adding a new kind of snapshot that is
specifically used for identifying if a given tuple, retrieved from the
database during an update or delete statement should be visible during the
current transaction.

This approach requires a new kind of snapshot (this idea was used by Gavin
for a previous updatable cursor patch but objections were raised.)

Option 3 MVCC Via Hash Table in Memory

The executor can be changed by saving into a hash table and comparing each
tuple in the cursor with that set to check if the tuple should be visible.
This approach has the advantage that it will be quick. It has the
disadvantage that, since the hash table will contain all the tuples of the
table being checked that it may use all local memory for a large table.

Option 4 MVCC Via Hash Table on Disk

When the UPDATE or DELETE request is executed the first time the Tid scan
database retrieval will be done first. At this time the tuple id of each row
in the table to be updated by the request will be available in the executor.
These tuple ids need to be stored in a hash table that is stored to disk, as,
if the table is large there could be a huge number of tuple ids. This data is
then available for comparison with the individual tuple to be updated or
deleted to check if it should be processed. The hash table will exist for the
duration of the transaction, from BEGIN to END (or ABORT).

The hash table is then used to identify if the tuple should be visible during
the current transaction. If the tuple should be visible then the update or
delete proceeds as usual.

This approach has the advantage that it will use little memory but will be
relatively slow as the data has to be accessed from disk.

Option 5 Store Tuple Id in Snapshot.

The Snapshot structure can be changed to include the tuple id. This enables
the current state of the tuple to be identified with respect to the current
transaction.
The tuple id, as identified in the cursor at the point where the
DELETE/UPDATE statement is being processed, can use the snapshot to identify
if the tuple should be visible in the context of the current transaction.

2.6 Changes to the Catalog
----------------------------
The Catalog needs to reflect changes introduced by the updatable cursor
implementation. A boolean attribute ï¿½is_for_updateï¿½ is to be added to the
pg_cursors implementation. It will define that the cursor is for update
(value is FALSE) or for share (value is TRUE, the default value).

3 Design Assumptions
----------------------------
The following design assumptions are made:

As PostgreSQL8.2 does not support the SENSITIVE cursor option the tuples
contained in a cursor can never be updated so these tuples will always appear
in their ï¿½originalï¿½ form as at the start of the transaction. This is in
breach of the SQL2003 Standard as described in 5WD-02-Foundation-2003-09.pdf,
p 810. The standard requires the updatable cursor to be declared as sensitive.

With respect to nested transactions ï¿½ In PostgreSQL nested transactions are
implemented by defining ï¿½save pointsï¿½ via the keyword SAVEPOINT. A ï¿½ROLLBACK
TO SAVEPOINTï¿½ rolls back the database contents to the last savepoint in this
transaction or the begin statement, whichever is closer.

It is assumed that the FETCH statement is used to return only a single row
into the cursor with each command when the cursor is updatable.

According to the SQL2003 Standard Update and Delete statements may contain
only a single base table.

The DECLARE CURSOR statement is supposed to use column level locking, but
PostgreSQL supports only row level locking. The result of this is that the
column list that the standard requires ï¿½DECLARE <cursor_name> SELECT ï¿½ FOR
UPDATE OF column-listï¿½ becomes a relation (table) list.

This is an email from Fujitsu Australia Software Technology Pty Ltd, ABN 27 003 693 481. It is confidential to the ordinary user of the email address to which it was addressed and may contain copyright and/or legally privileged information. No one else may read, print, store, copy or forward all or any of it or its attachments. If you receive this email in error, please return to sender. Thank you.

If you do not wish to receive commercial email messages from Fujitsu Australia Software Technology Pty Ltd, please email unsubscribe@fast.fujitsu.com.au

---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

Import Notes

Reply to msg id not found: 5537.10001169527168.fast.fujitsu.com.au@MHS

jd@commandprompt.com

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: Updateable cursors

FAST PostgreSQL wrote:

On Tue, 23 Jan 2007 15:48, Joshua D. Drake wrote:

FAST PostgreSQL wrote:

We are trying to develop the updateable cursors functionality into
Postgresql. I have given below details of the design and also issues we
are facing. Looking forward to the advice on how to proceed with these
issues.

Rgds,
Arul Shaji

Would this be something that you would hope to submit for 8.3?

Yes definitely. If we can finish it before the feature freeze of course.

Great! I will put it on my, "Remember to bug Arul" list :)

Sincerely,

Joshua D. Drake

Rgds,
Arul Shaji

Joshua D. Drake

1. Introduction
--------------
This is a combined proposal and design document for adding updatable
(insensitive) cursor capability to the PostgreSQL database.
There have already been a couple of previous proposals since 2003 for
implementing this feature so there appears to be community interest in
doing so. This will enable the following constructs to be processed:

UPDATE <table_name> SET value_list WHERE CURRENT OF <cursor_name>
DELETE FROM <table_name> WHERE CURRENT OF <cursor_name>

This has the effect of users being able to update or delete specific rows
of a table, as defined by the row currently fetched into the cursor.

2. Overall Conceptual Design
-----------------------------
The design is considered from the viewpoint of progression of a command
through the various stages of processing, from changes to the file
?gram.y? to implement the actual grammar changes, through to changes in
the Executor portion of the database architecture.

2.1 Changes to the Grammar
------------------------------
The following changes will be done to the PostgreSQL grammar:

UPDATE statement has the option ?WHERE CURRENT OF <cursor_name>? added
DELETE statement has the option ?WHERE CURRENT OF <cursor_name>? added

The cursor_name data is held in the UpdateStmt and DeleteStmt structures
and contains just the name of the cursor.

The pl/pgsql grammar changes in the same manner.

The word CURRENT will be added to the ScanKeywords array in keywords.c.

2.2 Changes to Affected Data Structures
------------------------------------------
The following data structures are affected by this change:

Portal structure, QueryDesc structure, the UpdateStmt and DeleteStmt
structures

The Portal will contain a list of structures of relation ids and tuple
ids relating to the tuple held in the QueryDesc structure. There will be
one entry in the relation and tuple id list for each entry in the
relation-list of the statement below:

DECLARE <cursor_name> [WITH HOLD] SELECT FOR UPDATE OF <relation-list>

The QueryDesc structure will contain the relation id and the tuple id
relating to the tuple obtained via the FETCH command so that it can be
propagated back to the Portal for storage in the list described above.

The UpdateStmt and DeleteStmt structures have the cursor name added so
that the information is available for use in obtaining the portal
structure related to the cursor previously opened via the DECLARE CURSOR
request.

2.3 Changes to the SQL Parser
------------------------------------
At present, although the FOR UPDATE clause of the DECLARE CURSOR command
has been present in the grammar, it causes an error message later in the
processing since cursors are currently not updatable. This now needs to
change. The ?FOR UPDATE? clause has to be valid, but not the ?FOR SHARE?
clause.

The relation names that follow the ?FOR UPDATE? clause will be added to
the rtable in the Query structure and identified by means of the rowMarks
array. In the case of an updatable cursor the FOR SHARE option is not
allowed therefore all entries in the rtable that are identified by the
rowMarks array must relate to tables that are FOR UPDATE.

In the UPDATE or DELETE statements the ?WHERE CURRENT OF <cursor_name>?
clause results in the cursor name being placed in the UpdateStmt or
DeleteStmt structure. During the processing of the functions -
transformDeleteStmt() and transformUpdateStmt() - the cursor name is used
to obtain a pointer to the related Portal structure and the tuple
affected by the current UPDATE or DELETE statement is extracted from the
Portal, where it has been placed as the result of a previous FETCH
request. At this point all the information for the UPDATE or DELETE
statement is available so the statements can be transformed into standard
UPDATE or DELETE statements and sent for re-write/planning/execution as
usual.

2.4 Changes to the Optimizer
------------------------------
There is a need to add a TidScan node to planning UPDATE / DELETE
statements where the statements are ?UPDATE / DELETE at position?. This
is to enable the tuple ids of the tuples in the tables relating to the
query to be obtained. There will need to be a new mechanism to achieve
this, as at present, a Tid scan is done only if there is a standard WHERE
condition on update or delete statements to provide Tid qualifier data.

2.5 Changes to the Executor
-------------------------------
There are various options that have been considered for this part of the
enhancement. These are described in the sections below.

We would like to hear opinions on which option is the best way to go or
if none of these is acceptable, any alternate ideas ?

Option 1 MVCC Via Continuous Searching of Database

The Executor is to be changed in the following ways:
1) When the FETCH statement is executed the id of the resulting tuple is
extracted and passed back to the Portal structure to be saved to indicate
the cursor is currently positioned on a tuple.
2) When the UPDATE or DELETE request is executed the tuple id previously
FETCHed is held in the QueryDesc structure so that it can be compared
with the tuple ids returned from the TidScan node processed prior to the
actual UPDATE / DELETE node in the plan. This enables a decision to be
made as to whether the tuple held in the cursor is visible to the UPDATE
/ DELETE request according to the rules of concurrency. The result is
that, at the cost of repeatedly searching the database at each UPDATE /
DELETE command, the hash table is no longer required.
This approach has the advantage that there is no hash table held in
memory or on disk so it will not be memory intensive but will be
processing intensive.

This is a good ?one-off? solution to the problem and, taken in isolation
is probably the best approach. However, if one considers the method(s)
used in other areas of PostgreSQL, it is probably not the best solution.
This option will probably not be used further.

Option 2 MVCC via New Snapshot

The executor can be changed by adding a new kind of snapshot that is
specifically used for identifying if a given tuple, retrieved from the
database during an update or delete statement should be visible during
the current transaction.

This approach requires a new kind of snapshot (this idea was used by
Gavin for a previous updatable cursor patch but objections were raised.)

Option 3 MVCC Via Hash Table in Memory

The executor can be changed by saving into a hash table and comparing
each tuple in the cursor with that set to check if the tuple should be
visible. This approach has the advantage that it will be quick. It has
the disadvantage that, since the hash table will contain all the tuples
of the table being checked that it may use all local memory for a large
table.

Option 4 MVCC Via Hash Table on Disk

When the UPDATE or DELETE request is executed the first time the Tid scan
database retrieval will be done first. At this time the tuple id of each
row in the table to be updated by the request will be available in the
executor. These tuple ids need to be stored in a hash table that is
stored to disk, as, if the table is large there could be a huge number of
tuple ids. This data is then available for comparison with the individual
tuple to be updated or deleted to check if it should be processed. The
hash table will exist for the duration of the transaction, from BEGIN to
END (or ABORT).

The hash table is then used to identify if the tuple should be visible
during the current transaction. If the tuple should be visible then the
update or delete proceeds as usual.

This approach has the advantage that it will use little memory but will
be relatively slow as the data has to be accessed from disk.

Option 5 Store Tuple Id in Snapshot.

The Snapshot structure can be changed to include the tuple id. This
enables the current state of the tuple to be identified with respect to
the current transaction.
The tuple id, as identified in the cursor at the point where the
DELETE/UPDATE statement is being processed, can use the snapshot to
identify if the tuple should be visible in the context of the current
transaction.

2.6 Changes to the Catalog
----------------------------
The Catalog needs to reflect changes introduced by the updatable cursor
implementation. A boolean attribute ?is_for_update? is to be added to the
pg_cursors implementation. It will define that the cursor is for update
(value is FALSE) or for share (value is TRUE, the default value).

3 Design Assumptions
----------------------------
The following design assumptions are made:

As PostgreSQL8.2 does not support the SENSITIVE cursor option the tuples
contained in a cursor can never be updated so these tuples will always
appear in their ?original? form as at the start of the transaction. This
is in breach of the SQL2003 Standard as described in
5WD-02-Foundation-2003-09.pdf, p 810. The standard requires the updatable
cursor to be declared as sensitive.

With respect to nested transactions ? In PostgreSQL nested transactions
are implemented by defining ?save points? via the keyword SAVEPOINT. A
?ROLLBACK TO SAVEPOINT? rolls back the database contents to the last
savepoint in this transaction or the begin statement, whichever is
closer.

It is assumed that the FETCH statement is used to return only a single
row into the cursor with each command when the cursor is updatable.

According to the SQL2003 Standard Update and Delete statements may
contain only a single base table.

The DECLARE CURSOR statement is supposed to use column level locking, but
PostgreSQL supports only row level locking. The result of this is that
the column list that the standard requires ?DECLARE <cursor_name> SELECT
? FOR UPDATE OF column-list? becomes a relation (table) list.

This is an email from Fujitsu Australia Software Technology Pty Ltd, ABN
27 003 693 481. It is confidential to the ordinary user of the email
address to which it was addressed and may contain copyright and/or
legally privileged information. No one else may read, print, store, copy
or forward all or any of it or its attachments. If you receive this email
in error, please return to sender. Thank you.

If you do not wish to receive commercial email messages from Fujitsu
Australia Software Technology Pty Ltd, please email
unsubscribe@fast.fujitsu.com.au

---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match

This is an email from Fujitsu Australia Software Technology Pty Ltd, ABN 27 003 693 481. It is confidential to the ordinary user of the email address to which it was addressed and may contain copyright and/or legally privileged information. No one else may read, print, store, copy or forward all or any of it or its attachments. If you receive this email in error, please return to sender. Thank you.

If you do not wish to receive commercial email messages from Fujitsu Australia Software Technology Pty Ltd, please email unsubscribe@fast.fujitsu.com.au

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

Import Notes

Reply to msg id not found: 5537.10011169527796.fast.fujitsu.com.au@MHS

Lukas Kahwe Smith

smith@pooteeweet.org

almost 19 years ago

In reply to: Joshua D. Drake (#8)

Re: Updateable cursors

Joshua D. Drake wrote:

Great! I will put it on my, "Remember to bug Arul" list :)

Hey Joshua,

could you put this stuff here:
http://developer.postgresql.org/index.php/Todo:WishlistFor83

I will try to find some time during this week (likely on the weekend) to
also try and figure out if these items are real and if the people still
think they can do them for 8.3 .. your additions would be most helpful.

regards,
Lukas

#10

jd@commandprompt.com

almost 19 years ago

In reply to: Lukas Kahwe Smith (#9)

Re: Updateable cursors

Lukas Kahwe Smith wrote:

Joshua D. Drake wrote:

Great! I will put it on my, "Remember to bug Arul" list :)

Hey Joshua,

could you put this stuff here:
http://developer.postgresql.org/index.php/Todo:WishlistFor83

Sure if you bother to unlock the page for me ;)

I will try to find some time during this week (likely on the weekend) to
also try and figure out if these items are real and if the people still
think they can do them for 8.3 .. your additions would be most helpful.

regards,
Lukas

---------------------------(end of broadcast)---------------------------
TIP 3: Have you checked our extensive FAQ?

http://www.postgresql.org/docs/faq

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#11

Heikki Linnakangas

heikki@enterprisedb.com

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: 10 weeks to feature freeze (Pending Work)

Joshua D. Drake wrote:

Or so... :)

Thought I would do a poll of what is happening in the world for 8.3. I have:

Alvaro Herrera: Autovacuum improvements (maintenance window etc..)
Gavin Sherry: Bitmap Indexes (on disk), possible basic Window functions

Gavin: how's it going with the bitmap indexes? I could work on it as
well, but I don't want to step on your toes.

Heikki Linnakangas: Working on Vacuum for Bitmap Indexes?

Yeah, that's the plan.

Also:
* Grouped Index Tuples (http://community.enterprisedb.com/git/). I don't
know how to proceed with this, but it's a feature I'd like to get in
8.3. Suggestions, anyone? I haven't received much comments on the design
or code...

* vacuum enhancements, not sure what exactly..

* Plan invalidation, possibly. Tom had plans on this as well.

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com

#12

Pavan Deolasee

pavan.deolasee@gmail.com

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: 10 weeks to feature freeze (Pending Work)

On 1/23/07, Joshua D. Drake <jd@commandprompt.com> wrote:

Or so... :)

I am sure there are more, the ones with question marks are unknowns but
heard of in the ether somewhere. Any additions or confirmations?

I have the first phase of Frequent Update Optimizations (HOT) patch ready.
But I held it back because of the concerns that its too complex. It has
shown decent performance gains on pgbench and DBT2 tests though.

I am splitting the patch into smaller pieces for ease of review and would
submit those soon for comments.

Thanks,
Pavan

EnterpriseDB http://www.enterprisedb.com

#13

Magnus Hagander

magnus@hagander.net

almost 19 years ago

In reply to: Stephen Frost (#6)

Re: 10 weeks to feature freeze (Pending Work)

On Mon, Jan 22, 2007 at 09:14:01PM -0500, Stephen Frost wrote:

* Joshua D. Drake (jd@commandprompt.com) wrote:

Thought I would do a poll of what is happening in the world for 8.3. I have:

Another thing which was mentioned previously which I'd really like to
see happen (and was discussed on the list...) is replacing the Kerberos
support with GSSAPI support and adding support for SSPI. Don't recall
who had said they were looking into working on it though..

That's Henry B. Hotz. He's done some work on it, and I have some stuff
to comment on sitting in my mailbox that I haven't had time to look at
yet. But I'm going to try to do that soon so he can continue.

//Magnus

#14

simon@2ndquadrant.com

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: Updateable cursors

On Wed, 2007-01-24 at 02:42 +1100, FAST PostgreSQL wrote:

In the UPDATE or DELETE statements the ‘WHERE CURRENT OF <cursor_name>’
clause results in the cursor name being placed in the UpdateStmt or
DeleteStmt structure. During the processing of the functions -
transformDeleteStmt() and transformUpdateStmt() - the cursor name is used to
obtain a pointer to the related Portal structure

To support prepared statements we'd need to do this name lookup just
once, so that the Update/Delete stmt can record which Portal to look at
for the current tuple.

and the tuple affected by
the current UPDATE or DELETE statement is extracted from the Portal, where it
has been placed as the result of a previous FETCH request. At this point all
the information for the UPDATE or DELETE statement is available so the
statements can be transformed into standard UPDATE or DELETE statements and
sent for re-write/planning/execution as usual.

2.5 Changes to the Executor
-------------------------------
There are various options that have been considered for this part of the
enhancement. These are described in the sections below.

Option 1 MVCC Via Continuous Searching of Database

The Executor is to be changed in the following ways:
1) When the FETCH statement is executed the id of the resulting tuple is
extracted and passed back to the Portal structure to be saved to indicate the
cursor is currently positioned on a tuple.
2) When the UPDATE or DELETE request is executed the tuple id previously
FETCHed is held in the QueryDesc structure so that it can be compared with
the tuple ids returned from the TidScan node processed prior to the actual
UPDATE / DELETE node in the plan. This enables a decision to be made as to
whether the tuple held in the cursor is visible to the UPDATE / DELETE
request according to the rules of concurrency. The result is that, at the
cost of repeatedly searching the database at each UPDATE / DELETE command,
the hash table is no longer required.
This approach has the advantage that there is no hash table held in memory or
on disk so it will not be memory intensive but will be processing intensive.

Do you have a specific example that would cause problems? It's much
easier to give examples that might cause problems and discuss those.

AFAICS in the straightforward case the Fetch will only return rows it
can see so update/delete should have no problems, iff the update/delete
is using a same or later snapshot than the cursor. I can see potential
problems with scrollable cursors.

So I'm not sure why there's a big need for any of the 5 options, yet.

--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com

Import Notes

Reply to msg id not found: 5537.10001169527168.fast.fujitsu.com.au@MHS

#15

Lukas Kahwe Smith

smith@pooteeweet.org

almost 19 years ago

In reply to: Joshua D. Drake (#10)

Re: Updateable cursors

Joshua D. Drake wrote:

Lukas Kahwe Smith wrote:

Joshua D. Drake wrote:

Great! I will put it on my, "Remember to bug Arul" list :)

Hey Joshua,

could you put this stuff here:
http://developer.postgresql.org/index.php/Todo:WishlistFor83

Sure if you bother to unlock the page for me ;)

hmm .. i am not aware of having a lock. i dont know mediawiki all that
well, but clicking around i could not find anything. IIRC someone else
also had issues editing pages on the wiki.

regards,
Lukas

#16

Iannsp

iannsp@gmail.com

almost 19 years ago

In reply to: Pavan Deolasee (#12)

About PostgreSQL certification

Hello,
I did like to know what you think about the postgresql certifications
provided for

PostgreSQL CE
http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst
http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the clients.
Make difference to be certified?

thanks for advanced.

Ivo Nascimento.

#17

Teodor Sigaev

teodor@sigaev.ru

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: 10 weeks to feature freeze (Pending Work)

I would like to suggest patches for OR-clause optimization and using index for
searching NULLs.

--
Teodor Sigaev E-mail: teodor@sigaev.ru
WWW: http://www.sigaev.ru/

#18

Tom Lane

tgl@sss.pgh.pa.us

almost 19 years ago

In reply to: Simon Riggs (#14)

Re: Updateable cursors

"Simon Riggs" <simon@2ndquadrant.com> writes:

On Wed, 2007-01-24 at 02:42 +1100, FAST PostgreSQL wrote:

In the UPDATE or DELETE statements the ‘WHERE CURRENT OF <cursor_name>’
clause results in the cursor name being placed in the UpdateStmt or
DeleteStmt structure. During the processing of the functions -
transformDeleteStmt() and transformUpdateStmt() - the cursor name is used to
obtain a pointer to the related Portal structure

To support prepared statements we'd need to do this name lookup just
once, so that the Update/Delete stmt can record which Portal to look at
for the current tuple.

This really isn't gonna work, because it assumes that the tuple that is
"current" at the instant of parsing is still going to be "current" at
execution time.

regards, tom lane

#19

simon@2ndquadrant.com

almost 19 years ago

In reply to: Tom Lane (#18)

Re: Updateable cursors

On Tue, 2007-01-23 at 09:55 -0500, Tom Lane wrote:

"Simon Riggs" <simon@2ndquadrant.com> writes:

On Wed, 2007-01-24 at 02:42 +1100, FAST PostgreSQL wrote:

In the UPDATE or DELETE statements the ‘WHERE CURRENT OF <cursor_name>’
clause results in the cursor name being placed in the UpdateStmt or
DeleteStmt structure. During the processing of the functions -
transformDeleteStmt() and transformUpdateStmt() - the cursor name is used to
obtain a pointer to the related Portal structure

To support prepared statements we'd need to do this name lookup just
once, so that the Update/Delete stmt can record which Portal to look at
for the current tuple.

This really isn't gonna work, because it assumes that the tuple that is
"current" at the instant of parsing is still going to be "current" at
execution time.

Of course thats true, but you've misread my comment.

The portal with the cursor in will not change, no matter how many times
we execute WHERE CURRENT OF in another portal. The OP suggested putting
the current tuple pointer onto the portal data, so this will work.

--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com

#20

Tom Lane

tgl@sss.pgh.pa.us

almost 19 years ago

In reply to: Simon Riggs (#19)

Re: Updateable cursors

"Simon Riggs" <simon@2ndquadrant.com> writes:

On Tue, 2007-01-23 at 09:55 -0500, Tom Lane wrote:

This really isn't gonna work, because it assumes that the tuple that is
"current" at the instant of parsing is still going to be "current" at
execution time.

Of course thats true, but you've misread my comment.

The portal with the cursor in will not change, no matter how many times
we execute WHERE CURRENT OF in another portal.

Really? The cursor portal will cease to exist as soon as the
transaction ends, but the prepared plan won't. A reasonable person
would expect that WHERE CURRENT OF will parse into a plan that just
stores the cursor name, and looks up the cursor at execution time.

The OP suggested putting
the current tuple pointer onto the portal data, so this will work.

No, as I read his message he was suggesting pulling data out of the
cursor portal at plan time so that no downstream (executor) changes
would be needed. That is certainly never going to be workable.

regards, tom lane

#21

FAST PostgreSQL

fastpgs@fast.fujitsu.com.au

almost 19 years ago

In reply to: Stephen Frost (#6)

Updateable cursors

We are trying to develop the updateable cursors functionality into
Postgresql. I have given below details of the design and also issues we are
facing. Looking forward to the advice on how to proceed with these issues.

Rgds,
Arul Shaji

1. Introduction
--------------
This is a combined proposal and design document for adding updatable
(insensitive) cursor capability to the PostgreSQL database.
There have already been a couple of previous proposals since 2003 for
implementing this feature so there appears to be community interest in doing
so. This will enable the following constructs to be processed:

UPDATE <table_name> SET value_list WHERE CURRENT OF <cursor_name>
DELETE FROM <table_name> WHERE CURRENT OF <cursor_name>

This has the effect of users being able to update or delete specific rows of
a table, as defined by the row currently fetched into the cursor.

2. Overall Conceptual Design
-----------------------------
The design is considered from the viewpoint of progression of a command
through the various stages of processing, from changes to the file ï¿½gram.yï¿½
to implement the actual grammar changes, through to changes in the Executor
portion of the database architecture.

2.1 Changes to the Grammar
------------------------------
The following changes will be done to the PostgreSQL grammar:

UPDATE statement has the option ï¿½WHERE CURRENT OF <cursor_name>ï¿½ added
DELETE statement has the option ï¿½WHERE CURRENT OF <cursor_name>ï¿½ added

The cursor_name data is held in the UpdateStmt and DeleteStmt structures and
contains just the name of the cursor.

The pl/pgsql grammar changes in the same manner.

The word CURRENT will be added to the ScanKeywords array in keywords.c.

2.2 Changes to Affected Data Structures
------------------------------------------
The following data structures are affected by this change:

Portal structure, QueryDesc structure, the UpdateStmt and DeleteStmt
structures

The Portal will contain a list of structures of relation ids and tuple ids
relating to the tuple held in the QueryDesc structure. There will be one
entry in the relation and tuple id list for each entry in the relation-list
of the statement below:

DECLARE <cursor_name> [WITH HOLD] SELECT FOR UPDATE OF <relation-list>

The QueryDesc structure will contain the relation id and the tuple id
relating to the tuple obtained via the FETCH command so that it can be
propagated back to the Portal for storage in the list described above.

The UpdateStmt and DeleteStmt structures have the cursor name added so that
the information is available for use in obtaining the portal structure
related to the cursor previously opened via the DECLARE CURSOR request.

2.3 Changes to the SQL Parser
------------------------------------
At present, although the FOR UPDATE clause of the DECLARE CURSOR command has
been present in the grammar, it causes an error message later in the
processing since cursors are currently not updatable. This now needs to
change. The ï¿½FOR UPDATEï¿½ clause has to be valid, but not the ï¿½FOR SHAREï¿½
clause.

The relation names that follow the ï¿½FOR UPDATEï¿½ clause will be added to the
rtable in the Query structure and identified by means of the rowMarks array.
In the case of an updatable cursor the FOR SHARE option is not allowed
therefore all entries in the rtable that are identified by the rowMarks array
must relate to tables that are FOR UPDATE.

In the UPDATE or DELETE statements the ï¿½WHERE CURRENT OF <cursor_name>ï¿½
clause results in the cursor name being placed in the UpdateStmt or
DeleteStmt structure. During the processing of the functions -
transformDeleteStmt() and transformUpdateStmt() - the cursor name is used to
obtain a pointer to the related Portal structure and the tuple affected by
the current UPDATE or DELETE statement is extracted from the Portal, where it
has been placed as the result of a previous FETCH request. At this point all
the information for the UPDATE or DELETE statement is available so the
statements can be transformed into standard UPDATE or DELETE statements and
sent for re-write/planning/execution as usual.

2.4 Changes to the Optimizer
------------------------------
There is a need to add a TidScan node to planning UPDATE / DELETE statements
where the statements are ï¿½UPDATE / DELETE at positionï¿½. This is to enable the
tuple ids of the tuples in the tables relating to the query to be obtained.
There will need to be a new mechanism to achieve this, as at present, a Tid
scan is done only if there is a standard WHERE condition on update or delete
statements to provide Tid qualifier data.

2.5 Changes to the Executor
-------------------------------
There are various options that have been considered for this part of the
enhancement. These are described in the sections below.

We would like to hear opinions on which option is the best way to go or if
none of these is acceptable, any alternate ideas ?

Option 1 MVCC Via Continuous Searching of Database

The Executor is to be changed in the following ways:
1) When the FETCH statement is executed the id of the resulting tuple is
extracted and passed back to the Portal structure to be saved to indicate the
cursor is currently positioned on a tuple.
2) When the UPDATE or DELETE request is executed the tuple id previously
FETCHed is held in the QueryDesc structure so that it can be compared with
the tuple ids returned from the TidScan node processed prior to the actual
UPDATE / DELETE node in the plan. This enables a decision to be made as to
whether the tuple held in the cursor is visible to the UPDATE / DELETE
request according to the rules of concurrency. The result is that, at the
cost of repeatedly searching the database at each UPDATE / DELETE command,
the hash table is no longer required.
This approach has the advantage that there is no hash table held in memory or
on disk so it will not be memory intensive but will be processing intensive.

This is a good ï¿½one-offï¿½ solution to the problem and, taken in isolation is
probably the best approach. However, if one considers the method(s) used in
other areas of PostgreSQL, it is probably not the best solution. This option
will probably not be used further.

Option 2 MVCC via New Snapshot

The executor can be changed by adding a new kind of snapshot that is
specifically used for identifying if a given tuple, retrieved from the
database during an update or delete statement should be visible during the
current transaction.

This approach requires a new kind of snapshot (this idea was used by Gavin
for a previous updatable cursor patch but objections were raised.)

Option 3 MVCC Via Hash Table in Memory

The executor can be changed by saving into a hash table and comparing each
tuple in the cursor with that set to check if the tuple should be visible.
This approach has the advantage that it will be quick. It has the
disadvantage that, since the hash table will contain all the tuples of the
table being checked that it may use all local memory for a large table.

Option 4 MVCC Via Hash Table on Disk

When the UPDATE or DELETE request is executed the first time the Tid scan
database retrieval will be done first. At this time the tuple id of each row
in the table to be updated by the request will be available in the executor.
These tuple ids need to be stored in a hash table that is stored to disk, as,
if the table is large there could be a huge number of tuple ids. This data is
then available for comparison with the individual tuple to be updated or
deleted to check if it should be processed. The hash table will exist for the
duration of the transaction, from BEGIN to END (or ABORT).

The hash table is then used to identify if the tuple should be visible during
the current transaction. If the tuple should be visible then the update or
delete proceeds as usual.

This approach has the advantage that it will use little memory but will be
relatively slow as the data has to be accessed from disk.

Option 5 Store Tuple Id in Snapshot.

The Snapshot structure can be changed to include the tuple id. This enables
the current state of the tuple to be identified with respect to the current
transaction.
The tuple id, as identified in the cursor at the point where the
DELETE/UPDATE statement is being processed, can use the snapshot to identify
if the tuple should be visible in the context of the current transaction.

2.6 Changes to the Catalog
----------------------------
The Catalog needs to reflect changes introduced by the updatable cursor
implementation. A boolean attribute ï¿½is_for_updateï¿½ is to be added to the
pg_cursors implementation. It will define that the cursor is for update
(value is FALSE) or for share (value is TRUE, the default value).

3 Design Assumptions
----------------------------
The following design assumptions are made:

As PostgreSQL8.2 does not support the SENSITIVE cursor option the tuples
contained in a cursor can never be updated so these tuples will always appear
in their ï¿½originalï¿½ form as at the start of the transaction. This is in
breach of the SQL2003 Standard as described in 5WD-02-Foundation-2003-09.pdf,
p 810. The standard requires the updatable cursor to be declared as sensitive.

With respect to nested transactions ï¿½ In PostgreSQL nested transactions are
implemented by defining ï¿½save pointsï¿½ via the keyword SAVEPOINT. A ï¿½ROLLBACK
TO SAVEPOINTï¿½ rolls back the database contents to the last savepoint in this
transaction or the begin statement, whichever is closer.

It is assumed that the FETCH statement is used to return only a single row
into the cursor with each command when the cursor is updatable.

According to the SQL2003 Standard Update and Delete statements may contain
only a single base table.

The DECLARE CURSOR statement is supposed to use column level locking, but
PostgreSQL supports only row level locking. The result of this is that the
column list that the standard requires ï¿½DECLARE <cursor_name> SELECT ï¿½ FOR
UPDATE OF column-listï¿½ becomes a relation (table) list.

This is an email from Fujitsu Australia Software Technology Pty Ltd, ABN 27 003 693 481. It is confidential to the ordinary user of the email address to which it was addressed and may contain copyright and/or legally privileged information. No one else may read, print, store, copy or forward all or any of it or its attachments. If you receive this email in error, please return to sender. Thank you.

If you do not wish to receive commercial email messages from Fujitsu Australia Software Technology Pty Ltd, please email unsubscribe@fast.fujitsu.com.au

#22

FAST PostgreSQL

fastpgs@fast.fujitsu.com.au

almost 19 years ago

In reply to: Joshua D. Drake (#7)

Re: Updateable cursors

On Tue, 23 Jan 2007 15:48, Joshua D. Drake wrote:

FAST PostgreSQL wrote:

We are trying to develop the updateable cursors functionality into
Postgresql. I have given below details of the design and also issues we
are facing. Looking forward to the advice on how to proceed with these
issues.

Rgds,
Arul Shaji

Would this be something that you would hope to submit for 8.3?

Yes definitely. If we can finish it before the feature freeze of course.

Rgds,
Arul Shaji

Joshua D. Drake

1. Introduction
--------------
This is a combined proposal and design document for adding updatable
(insensitive) cursor capability to the PostgreSQL database.
There have already been a couple of previous proposals since 2003 for
implementing this feature so there appears to be community interest in
doing so. This will enable the following constructs to be processed:

UPDATE <table_name> SET value_list WHERE CURRENT OF <cursor_name>
DELETE FROM <table_name> WHERE CURRENT OF <cursor_name>

This has the effect of users being able to update or delete specific rows
of a table, as defined by the row currently fetched into the cursor.

2. Overall Conceptual Design
-----------------------------
The design is considered from the viewpoint of progression of a command
through the various stages of processing, from changes to the file
?gram.y? to implement the actual grammar changes, through to changes in
the Executor portion of the database architecture.

2.1 Changes to the Grammar
------------------------------
The following changes will be done to the PostgreSQL grammar:

UPDATE statement has the option ?WHERE CURRENT OF <cursor_name>? added
DELETE statement has the option ?WHERE CURRENT OF <cursor_name>? added

The cursor_name data is held in the UpdateStmt and DeleteStmt structures
and contains just the name of the cursor.

The pl/pgsql grammar changes in the same manner.

The word CURRENT will be added to the ScanKeywords array in keywords.c.

2.2 Changes to Affected Data Structures
------------------------------------------
The following data structures are affected by this change:

Portal structure, QueryDesc structure, the UpdateStmt and DeleteStmt
structures

The Portal will contain a list of structures of relation ids and tuple
ids relating to the tuple held in the QueryDesc structure. There will be
one entry in the relation and tuple id list for each entry in the
relation-list of the statement below:

DECLARE <cursor_name> [WITH HOLD] SELECT FOR UPDATE OF <relation-list>

The QueryDesc structure will contain the relation id and the tuple id
relating to the tuple obtained via the FETCH command so that it can be
propagated back to the Portal for storage in the list described above.

The UpdateStmt and DeleteStmt structures have the cursor name added so
that the information is available for use in obtaining the portal
structure related to the cursor previously opened via the DECLARE CURSOR
request.

2.3 Changes to the SQL Parser
------------------------------------
At present, although the FOR UPDATE clause of the DECLARE CURSOR command
has been present in the grammar, it causes an error message later in the
processing since cursors are currently not updatable. This now needs to
change. The ?FOR UPDATE? clause has to be valid, but not the ?FOR SHARE?
clause.

The relation names that follow the ?FOR UPDATE? clause will be added to
the rtable in the Query structure and identified by means of the rowMarks
array. In the case of an updatable cursor the FOR SHARE option is not
allowed therefore all entries in the rtable that are identified by the
rowMarks array must relate to tables that are FOR UPDATE.

In the UPDATE or DELETE statements the ?WHERE CURRENT OF <cursor_name>?
clause results in the cursor name being placed in the UpdateStmt or
DeleteStmt structure. During the processing of the functions -
transformDeleteStmt() and transformUpdateStmt() - the cursor name is used
to obtain a pointer to the related Portal structure and the tuple
affected by the current UPDATE or DELETE statement is extracted from the
Portal, where it has been placed as the result of a previous FETCH
request. At this point all the information for the UPDATE or DELETE
statement is available so the statements can be transformed into standard
UPDATE or DELETE statements and sent for re-write/planning/execution as
usual.

2.4 Changes to the Optimizer
------------------------------
There is a need to add a TidScan node to planning UPDATE / DELETE
statements where the statements are ?UPDATE / DELETE at position?. This
is to enable the tuple ids of the tuples in the tables relating to the
query to be obtained. There will need to be a new mechanism to achieve
this, as at present, a Tid scan is done only if there is a standard WHERE
condition on update or delete statements to provide Tid qualifier data.

2.5 Changes to the Executor
-------------------------------
There are various options that have been considered for this part of the
enhancement. These are described in the sections below.

We would like to hear opinions on which option is the best way to go or
if none of these is acceptable, any alternate ideas ?

Option 1 MVCC Via Continuous Searching of Database

The Executor is to be changed in the following ways:
1) When the FETCH statement is executed the id of the resulting tuple is
extracted and passed back to the Portal structure to be saved to indicate
the cursor is currently positioned on a tuple.
2) When the UPDATE or DELETE request is executed the tuple id previously
FETCHed is held in the QueryDesc structure so that it can be compared
with the tuple ids returned from the TidScan node processed prior to the
actual UPDATE / DELETE node in the plan. This enables a decision to be
made as to whether the tuple held in the cursor is visible to the UPDATE
/ DELETE request according to the rules of concurrency. The result is
that, at the cost of repeatedly searching the database at each UPDATE /
DELETE command, the hash table is no longer required.
This approach has the advantage that there is no hash table held in
memory or on disk so it will not be memory intensive but will be
processing intensive.

This is a good ?one-off? solution to the problem and, taken in isolation
is probably the best approach. However, if one considers the method(s)
used in other areas of PostgreSQL, it is probably not the best solution.
This option will probably not be used further.

Option 2 MVCC via New Snapshot

The executor can be changed by adding a new kind of snapshot that is
specifically used for identifying if a given tuple, retrieved from the
database during an update or delete statement should be visible during
the current transaction.

This approach requires a new kind of snapshot (this idea was used by
Gavin for a previous updatable cursor patch but objections were raised.)

Option 3 MVCC Via Hash Table in Memory

The executor can be changed by saving into a hash table and comparing
each tuple in the cursor with that set to check if the tuple should be
visible. This approach has the advantage that it will be quick. It has
the disadvantage that, since the hash table will contain all the tuples
of the table being checked that it may use all local memory for a large
table.

Option 4 MVCC Via Hash Table on Disk

When the UPDATE or DELETE request is executed the first time the Tid scan
database retrieval will be done first. At this time the tuple id of each
row in the table to be updated by the request will be available in the
executor. These tuple ids need to be stored in a hash table that is
stored to disk, as, if the table is large there could be a huge number of
tuple ids. This data is then available for comparison with the individual
tuple to be updated or deleted to check if it should be processed. The
hash table will exist for the duration of the transaction, from BEGIN to
END (or ABORT).

The hash table is then used to identify if the tuple should be visible
during the current transaction. If the tuple should be visible then the
update or delete proceeds as usual.

This approach has the advantage that it will use little memory but will
be relatively slow as the data has to be accessed from disk.

Option 5 Store Tuple Id in Snapshot.

The Snapshot structure can be changed to include the tuple id. This
enables the current state of the tuple to be identified with respect to
the current transaction.
The tuple id, as identified in the cursor at the point where the
DELETE/UPDATE statement is being processed, can use the snapshot to
identify if the tuple should be visible in the context of the current
transaction.

2.6 Changes to the Catalog
----------------------------
The Catalog needs to reflect changes introduced by the updatable cursor
implementation. A boolean attribute ?is_for_update? is to be added to the
pg_cursors implementation. It will define that the cursor is for update
(value is FALSE) or for share (value is TRUE, the default value).

3 Design Assumptions
----------------------------
The following design assumptions are made:

As PostgreSQL8.2 does not support the SENSITIVE cursor option the tuples
contained in a cursor can never be updated so these tuples will always
appear in their ?original? form as at the start of the transaction. This
is in breach of the SQL2003 Standard as described in
5WD-02-Foundation-2003-09.pdf, p 810. The standard requires the updatable
cursor to be declared as sensitive.

With respect to nested transactions ? In PostgreSQL nested transactions
are implemented by defining ?save points? via the keyword SAVEPOINT. A
?ROLLBACK TO SAVEPOINT? rolls back the database contents to the last
savepoint in this transaction or the begin statement, whichever is
closer.

It is assumed that the FETCH statement is used to return only a single
row into the cursor with each command when the cursor is updatable.

According to the SQL2003 Standard Update and Delete statements may
contain only a single base table.

The DECLARE CURSOR statement is supposed to use column level locking, but
PostgreSQL supports only row level locking. The result of this is that
the column list that the standard requires ?DECLARE <cursor_name> SELECT
? FOR UPDATE OF column-list? becomes a relation (table) list.

This is an email from Fujitsu Australia Software Technology Pty Ltd, ABN
27 003 693 481. It is confidential to the ordinary user of the email
address to which it was addressed and may contain copyright and/or
legally privileged information. No one else may read, print, store, copy
or forward all or any of it or its attachments. If you receive this email
in error, please return to sender. Thank you.

If you do not wish to receive commercial email messages from Fujitsu
Australia Software Technology Pty Ltd, please email
unsubscribe@fast.fujitsu.com.au

---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match

If you do not wish to receive commercial email messages from Fujitsu Australia Software Technology Pty Ltd, please email unsubscribe@fast.fujitsu.com.au

#23

jd@commandprompt.com

almost 19 years ago

In reply to: Pavan Deolasee (#12)

Re: 10 weeks to feature freeze (Pending Work)

Pavan Deolasee wrote:

On 1/23/07, Joshua D. Drake <jd@commandprompt.com> wrote:

Or so... :)

I am sure there are more, the ones with question marks are unknowns but
heard of in the ether somewhere. Any additions or confirmations?

I have the first phase of Frequent Update Optimizations (HOT) patch ready.
But I held it back because of the concerns that its too complex. It has
shown decent performance gains on pgbench and DBT2 tests though.

I am splitting the patch into smaller pieces for ease of review and would
submit those soon for comments.

*soon* is the operative word :).

Sincerely,

Joshua D. Drake

Thanks,
Pavan

EnterpriseDB http://www.enterprisedb.com

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#24

simon@2ndquadrant.com

almost 19 years ago

In reply to: Tom Lane (#20)

Re: Updateable cursors

On Tue, 2007-01-23 at 10:39 -0500, Tom Lane wrote:

"Simon Riggs" <simon@2ndquadrant.com> writes:

On Tue, 2007-01-23 at 09:55 -0500, Tom Lane wrote:

This really isn't gonna work, because it assumes that the tuple that is
"current" at the instant of parsing is still going to be "current" at
execution time.

Of course thats true, but you've misread my comment.

The portal with the cursor in will not change, no matter how many times
we execute WHERE CURRENT OF in another portal.

Really? The cursor portal will cease to exist as soon as the
transaction ends, but the prepared plan won't.

Yes, understood.

I just want it to work well with prepared queries also. That seems both
a reasonable goal and also achievable by caching in the way requested.

A reasonable person
would expect that WHERE CURRENT OF will parse into a plan that just
stores the cursor name, and looks up the cursor at execution time.

We just store the Xid for which the cache is relevant then refresh the
cache if the cache is stale.

If you don't like the idea, say so. There's no need for anything more.

But those are minor points if you have stronger reservations about the
main proposal, which it sounds like you do.

--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com

#25

Richard Troy

rtroy@ScienceTools.com

almost 19 years ago

In reply to: FAST PostgreSQL (#21)

Re: Updateable cursors

On Wed, 24 Jan 2007, FAST PostgreSQL wrote:

We are trying to develop the updateable cursors functionality into
Postgresql. I have given below details of the design and also issues we are
facing. Looking forward to the advice on how to proceed with these issues.

Rgds,
Arul Shaji

Hi Arul,

...I can see people are picking apart the implementation details so you're
getting good feedback on your ambitious proposal. Looks like you've put a
lot of thought/work into it.

I've never been a fan of cursors because they encourage bad behavior;
"Think time" in a transaction sometimes becomes "lunch time" for users and
in any event long lock duration is something to be avoided for the sake of
concurrency and sometimes performance (vacuum, etc). My philosophy is "get
in and get out quick."

Ten years ago May, our first customer insisted we implement what has
become our primary API library in Java and somewhat later I was shocked to
learn that for whatever reason Java ResultSets are supposed to be
implemented as _updateable_cursors._ This created serious security issues
for handing off results to other programs through the library - ones that
don't even have the ability to connect to the target database. Confirmed
in the behavior of Informix, we went through some hoops to remove the need
to pass ResultSets around. (If I had only known Postgres didn't implement
the RS as an updateable cursor, I'd have pushed for our primary platform
to be Postgres!)

What impresses me is that Postgres has survived so well without updateable
cursors. To my mind it illustrates that they aren't widely used. I'm
wondering what troubles lurk ahead once they're available. As a
DBA/SysAdmin, I'd be quite happy that there existed some kind of log
element that indicated updateable cursors were in use that I could search
for easily whenever trying to diagnose some performance or deadlocking
problem, etc, say log fiile entries that indicated the opening and later
closing of such a cursor with an id of some kind that allowed matching up
open/close pairs. I also think that that the documentation should be
updated to not only indicate usage of this new feature, but provide
cautionary warnings about the potential locking issues and, for the
authors of libraries, Java in particular, the possible security issues.

Regards,
Richard

--
Richard Troy, Chief Scientist
Science Tools Corporation
510-924-1363 or 202-747-1263
rtroy@ScienceTools.com, http://ScienceTools.com/

#26

Merlin Moncure

mmoncure@gmail.com

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: 10 weeks to feature freeze (Pending Work)

On 1/22/07, Joshua D. Drake <jd@commandprompt.com> wrote:

Or so... :)

Thought I would do a poll of what is happening in the world for 8.3. I have:

Alvaro Herrera: Autovacuum improvements (maintenance window etc..)
Gavin Sherry: Bitmap Indexes (on disk), possible basic Window functions
Jonah Harris: WITH/Recursive Queries?
Andrei Kovalesvki: Some Win32 work with Magnus
Magnus Hagander: VC++ support (thank goodness)
Heikki Linnakangas: Working on Vacuum for Bitmap Indexes?
Oleg Bartunov: Tsearch2 in core
Neil Conway: Patch Review (including enums), pg_fcache

has there been any progress on the 'hot' tuple update mechanism?

merlin

#27

jd@commandprompt.com

almost 19 years ago

In reply to: Iannsp (#16)

Re: About PostgreSQL certification

Iannsp wrote:

Hello,
I did like to know what you think about the postgresql certifications
provided for

PostgreSQL CE
http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst
http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the clients.
Make difference to be certified?'

It doesn't make a difference to be certified.

Sincerely,

Joshua D. Drake

thanks for advanced.

Ivo Nascimento.

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#28

David Fetter

david@fetter.org

almost 19 years ago

In reply to: Iannsp (#16)

Re: About PostgreSQL certification

On Tue, Jan 23, 2007 at 11:52:08AM -0200, Iannsp wrote:

Hello,
I did like to know what you think about the postgresql
certifications provided for

PostgreSQL CE http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the
clients. Make difference to be certified?

Clueful clients will look unfavorably on any "PostgreSQL
certification" you have. They will instead insist on experience and
references, as clueful clients do. :)

Cheers,
D
--
David Fetter <david@fetter.org> http://fetter.org/
phone: +1 415 235 3778 AIM: dfetter666
Skype: davidfetter

Remember to vote!

#29

jesus@omniti.com

almost 19 years ago

In reply to: David Fetter (#28)

Re: About PostgreSQL certification

On Jan 23, 2007, at 4:33 PM, David Fetter wrote:

On Tue, Jan 23, 2007 at 11:52:08AM -0200, Iannsp wrote:

Hello,
I did like to know what you think about the postgresql
certifications provided for

PostgreSQL CE http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the
clients. Make difference to be certified?

Clueful clients will look unfavorably on any "PostgreSQL
certification" you have. They will instead insist on experience and
references, as clueful clients do. :)

I don't believe that's true. Oracle certification means quite a
bit. Cisco certification is excellent. Sun certification is
decent. If the PostgreSQL certifications don't mean much it is a
problem with the particular vendor of the certificate and you (as a
PostgreSQL entity) should contest their right to use PostgreSQL name
in their advertising or marketing. Certification programs can and
should mean something.

We offer training programs here and have considered offering OmniTI
certifications in the future. I wouldn't offer then unless I thought
it meant something that companies "out there" could rely on. Many
other certifying entities have the same approach.

// Theo Schlossnagle
// CTO -- http://www.omniti.com/~jesus/
// OmniTI Computer Consulting, Inc. -- http://www.omniti.com/

#30

David Fetter

david@fetter.org

almost 19 years ago

In reply to: Theo Schlossnagle (#29)

Re: About PostgreSQL certification

On Tue, Jan 23, 2007 at 04:41:03PM -0500, Theo Schlossnagle wrote:

On Jan 23, 2007, at 4:33 PM, David Fetter wrote:

On Tue, Jan 23, 2007 at 11:52:08AM -0200, Iannsp wrote:

Hello,
I did like to know what you think about the postgresql
certifications provided for

PostgreSQL CE http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the
clients. Make difference to be certified?

Clueful clients will look unfavorably on any "PostgreSQL
certification" you have. They will instead insist on experience
and references, as clueful clients do. :)

I don't believe that's true. Oracle certification means quite a
bit. Cisco certification is excellent. Sun certification is
decent. If the PostgreSQL certifications don't mean much it is a
problem with the particular vendor of the certificate and you (as a
PostgreSQL entity) should contest their right to use PostgreSQL name
in their advertising or marketing.

Sadly, at least in the U.S., PostgreSQL is unlikely to be a defensible
trademark. I am not an intellectual property attorney, and if I were
one, my opinion would not be as weighty as a court case.

Certification programs can and should mean something.

I'd love to see a good one for PostgreSQL. What I've seen so far has
been somewhere between dismal and rotten.

We offer training programs here and have considered offering OmniTI
certifications in the future. I wouldn't offer then unless I
thought it meant something that companies "out there" could rely
on.

Great :)

Many other certifying entities have the same approach.

99% of them give the rest a bad name ;)

Cheers,
D
--
David Fetter <david@fetter.org> http://fetter.org/
phone: +1 415 235 3778 AIM: dfetter666
Skype: davidfetter

Remember to vote!

#31

Mark Kirkwood

markir@paradise.net.nz

almost 19 years ago

In reply to: Theo Schlossnagle (#29)

Re: About PostgreSQL certification

Theo Schlossnagle wrote:

On Jan 23, 2007, at 4:33 PM, David Fetter wrote:

On Tue, Jan 23, 2007 at 11:52:08AM -0200, Iannsp wrote:

Hello,
I did like to know what you think about the postgresql
certifications provided for

PostgreSQL CE http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the
clients. Make difference to be certified?

Clueful clients will look unfavorably on any "PostgreSQL
certification" you have. They will instead insist on experience and
references, as clueful clients do. :)

I don't believe that's true. Oracle certification means quite a bit.
Cisco certification is excellent. Sun certification is decent. If the
PostgreSQL certifications don't mean much it is a problem with the
particular vendor of the certificate and you (as a PostgreSQL entity)
should contest their right to use PostgreSQL name in their advertising
or marketing. Certification programs can and should mean something.

Certification is ok - but is only of actual value when combined with
real experience. The reason I say this is that certification programs in
general can be beaten by various techniques (e.g. friends, online
research, guessing etc). Also over time they are rendered (almost)
useless by the (lucrative) side businesses that come into being (e.g.
'boot camps', mock exams etc).

I think seeing relevant training courses + experience on a CV trumps
certification anytime - unfortunately a lot of folks out there are
mesmerized by shiny certificates....

Cheers

Mark

#32

jesus@omniti.com

almost 19 years ago

In reply to: Mark Kirkwood (#31)

Re: About PostgreSQL certification

On Jan 23, 2007, at 5:04 PM, Mark Kirkwood wrote:

Theo Schlossnagle wrote:

On Jan 23, 2007, at 4:33 PM, David Fetter wrote:

On Tue, Jan 23, 2007 at 11:52:08AM -0200, Iannsp wrote:

Hello,
I did like to know what you think about the postgresql
certifications provided for

PostgreSQL CE http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the
clients. Make difference to be certified?

Clueful clients will look unfavorably on any "PostgreSQL
certification" you have. They will instead insist on experience and
references, as clueful clients do. :)

I don't believe that's true. Oracle certification means quite a
bit. Cisco certification is excellent. Sun certification is
decent. If the PostgreSQL certifications don't mean much it is a
problem with the particular vendor of the certificate and you (as
a PostgreSQL entity) should contest their right to use PostgreSQL
name in their advertising or marketing. Certification programs
can and should mean something.

Certification is ok - but is only of actual value when combined
with real experience. The reason I say this is that certification
programs in general can be beaten by various techniques (e.g.
friends, online research, guessing etc). Also over time they are
rendered (almost) useless by the (lucrative) side businesses that
come into being (e.g. 'boot camps', mock exams etc).

Get a CCIE and tell me that again :-) When you are handed a
complicated network of routers and switches running all sorts of
version of IOS and CatOS and you go to lunch, they break it and you
have a certain time allotment to fix it all.

Most certifications are not simple multiple choice quizes. Just the
ones you hear about -- the ones that suck.

I think seeing relevant training courses + experience on a CV
trumps certification anytime - unfortunately a lot of folks out
there are mesmerized by shiny certificates....

Sure. But experience is very hard to get. And since people with
PostgreSQL experience are limited, companies adopting it need a good
second option -- certified people.

// Theo Schlossnagle
// CTO -- http://www.omniti.com/~jesus/
// OmniTI Computer Consulting, Inc. -- http://www.omniti.com/

#33

jd@commandprompt.com

almost 19 years ago

In reply to: Theo Schlossnagle (#29)

Re: About PostgreSQL certification

Theo Schlossnagle wrote:

On Jan 23, 2007, at 4:33 PM, David Fetter wrote:

On Tue, Jan 23, 2007 at 11:52:08AM -0200, Iannsp wrote:

Hello,
I did like to know what you think about the postgresql
certifications provided for

PostgreSQL CE http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the
clients. Make difference to be certified?

Clueful clients will look unfavorably on any "PostgreSQL
certification" you have. They will instead insist on experience and
references, as clueful clients do. :)

I don't believe that's true. Oracle certification means quite a bit.
Cisco certification is excellent. Sun certification is decent.

I agree that their are certifications that are worth something,
specifically those you mention above.

However, PostgreSQL certifications are not currently worth anything IMO.

Sincerely,

Joshua D. Drake
--

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#34

jd@commandprompt.com

almost 19 years ago

In reply to: Theo Schlossnagle (#32)

Re: About PostgreSQL certification

Get a CCIE and tell me that again :-) When you are handed a complicated
network of routers and switches running all sorts of version of IOS and
CatOS and you go to lunch, they break it and you have a certain time
allotment to fix it all.

Most certifications are not simple multiple choice quizes. Just the
ones you hear about -- the ones that suck.

I think seeing relevant training courses + experience on a CV trumps
certification anytime - unfortunately a lot of folks out there are
mesmerized by shiny certificates....

Sure. But experience is very hard to get. And since people with
PostgreSQL experience are limited, companies adopting it need a good
second option -- certified people.

They aren't limited, just all employed ;)

Joshua D. Drake

// Theo Schlossnagle
// CTO -- http://www.omniti.com/~jesus/
// OmniTI Computer Consulting, Inc. -- http://www.omniti.com/

---------------------------(end of broadcast)---------------------------
TIP 5: don't forget to increase your free space map settings

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#35

jesus@omniti.com

almost 19 years ago

In reply to: Joshua D. Drake (#34)

Re: About PostgreSQL certification

On Jan 23, 2007, at 5:14 PM, Joshua D. Drake wrote:

Get a CCIE and tell me that again :-) When you are handed a
complicated
network of routers and switches running all sorts of version of
IOS and
CatOS and you go to lunch, they break it and you have a certain time
allotment to fix it all.

Most certifications are not simple multiple choice quizes. Just the
ones you hear about -- the ones that suck.

I think seeing relevant training courses + experience on a CV trumps
certification anytime - unfortunately a lot of folks out there are
mesmerized by shiny certificates....

Sure. But experience is very hard to get. And since people with
PostgreSQL experience are limited, companies adopting it need a good
second option -- certified people.

They aren't limited, just all employed ;)

I can't find 500, let alone 1000, people with extensive postgresql
experience in an enterprise environment. Oracle has an order of
magnitude more. MySQL even has better numbers than postgres in this
arena. If you only want to hire people with extensive experience,
you're exposing yourself to an enormous business risk by adopting
postgres. You'd have to hire out to a consulting company and if too
many do that, the consulting company will have scaling issues (as all
do).

The upside of Oracle is that I can hire out to a consulting company
for some things (particularly challenging scale or recovery issues)
and get someone who knows their way around Oracle reasonably well
(has performed _real_ disaster recovery in a hands on fashion,
performed hands-on query tuning, database sizing exercises, etc.) by
simply finding someone who is Oracle certified (all of those things
are part of the Oracle certification process). Granted, just because
someone is certified doesn't mean they "fit" or will excel at the
problems you give them -- it's just a nice lower bar. Granted you
can make a name for yourself as an expert without getting a
certification, but if you've made a name for yourself, you aren't
likely to be on the job market -- which is really my point. Oracle's
certification programs have helped Oracle considerably in gaining the
number of Oracle professionals in the job market. PostgreSQL
certification has the opportunity to do the same and in doing so
increase overall PostgreSQL adoption. That's a good thing.

--
Theo

// Theo Schlossnagle
// CTO -- http://www.omniti.com/~jesus/
// OmniTI Computer Consulting, Inc. -- http://www.omniti.com/

#36

David Fetter

david@fetter.org

almost 19 years ago

In reply to: Theo Schlossnagle (#35)

Re: About PostgreSQL certification

On Tue, Jan 23, 2007 at 05:19:45PM -0500, Theo Schlossnagle wrote:

On Jan 23, 2007, at 5:14 PM, Joshua D. Drake wrote:

Get a CCIE and tell me that again :-) When you are handed a
complicated network of routers and switches running all sorts of
version of IOS and CatOS and you go to lunch, they break it and
you have a certain time allotment to fix it all.

Most certifications are not simple multiple choice quizes. Just
the ones you hear about -- the ones that suck.

I think seeing relevant training courses + experience on a CV
trumps certification anytime - unfortunately a lot of folks out
there are mesmerized by shiny certificates....

Sure. But experience is very hard to get. And since people with
PostgreSQL experience are limited, companies adopting it need a
good second option -- certified people.

They aren't limited, just all employed ;)

I can't find 500, let alone 1000, people with extensive postgresql
experience in an enterprise environment. Oracle has an order of
magnitude more. MySQL even has better numbers than postgres in this
arena. If you only want to hire people with extensive experience,
you're exposing yourself to an enormous business risk by adopting
postgres. You'd have to hire out to a consulting company and if too
many do that, the consulting company will have scaling issues (as
all do).

The upside of Oracle is that I can hire out to a consulting company
for some things (particularly challenging scale or recovery issues)
and get someone who knows their way around Oracle reasonably well
(has performed _real_ disaster recovery in a hands on fashion,
performed hands-on query tuning, database sizing exercises, etc.) by
simply finding someone who is Oracle certified (all of those things
are part of the Oracle certification process). Granted, just
because someone is certified doesn't mean they "fit" or will excel
at the problems you give them -- it's just a nice lower bar.
Granted you can make a name for yourself as an expert without
getting a certification, but if you've made a name for yourself,
you aren't likely to be on the job market -- which is really my
point. Oracle's certification programs have helped Oracle
considerably in gaining the number of Oracle professionals in the
job market. PostgreSQL certification has the opportunity to do the
same and in doing so increase overall PostgreSQL adoption. That's
a good thing.

When you're getting this together, by all means let me know so I can
trumpet it all over the PostgreSQL Weekly News :)

Cheers,
D
--
David Fetter <david@fetter.org> http://fetter.org/
phone: +1 415 235 3778 AIM: dfetter666
Skype: davidfetter

Remember to vote!

#37

jd@commandprompt.com

almost 19 years ago

In reply to: Theo Schlossnagle (#35)

Re: About PostgreSQL certification

Oracle's certification programs have helped Oracle
considerably in gaining the number of Oracle professionals in the job
market. PostgreSQL certification has the opportunity to do the same and
in doing so increase overall PostgreSQL adoption. That's a good thing.

Well maybe it is just me, but I am perfectly happy with PostgreSQL's
growth. I find that our customers are of a much higher quality than your
average MySQL or Oracle customer.

Sincerely,

Joshua D. Drake

--
Theo

// Theo Schlossnagle
// CTO -- http://www.omniti.com/~jesus/
// OmniTI Computer Consulting, Inc. -- http://www.omniti.com/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#38

Mark Kirkwood

markir@paradise.net.nz

almost 19 years ago

In reply to: Theo Schlossnagle (#32)

Re: About PostgreSQL certification

Theo Schlossnagle wrote:

Get a CCIE and tell me that again :-) When you are handed a complicated
network of routers and switches running all sorts of version of IOS and
CatOS and you go to lunch, they break it and you have a certain time
allotment to fix it all.

I know all about CCIE - one session fixing up hardware is no substitute
for experience (and is still vulnerable to the methods I mentioned - it
is however a lot better than a multi choice exam of course).

To cure the shortage of experienced Postgres folks there is only one
solution - err, more experience! So the need is for good training
courses (not necessarily certification and all the IMHO nonsense that
comes with that), and a willingness on the part of employers to invest
in upskilling their staff.

Cheers

Mark

#39

jd@commandprompt.com

almost 19 years ago

In reply to: Mark Kirkwood (#38)

Re: About PostgreSQL certification

To cure the shortage of experienced Postgres folks there is only one
solution - err, more experience! So the need is for good training
courses (not necessarily certification and all the IMHO nonsense that
comes with that), and a willingness on the part of employers to invest
in upskilling their staff.

You know its funny. Command Prompt, OTG-Inc, SRA and Big Nerd Ranch
*all* offer training.

Last I checked, OTG had to cancel classes because of lack of demand
(please verify Chander).

Command Prompt currently only trains corps with 12+ people per class so
we are a bit different.

Sinerely,

Joshua D. Drake

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#40

Mark Kirkwood

markir@paradise.net.nz

almost 19 years ago

In reply to: Joshua D. Drake (#39)

Re: About PostgreSQL certification

Joshua D. Drake wrote:

To cure the shortage of experienced Postgres folks there is only one
solution - err, more experience! So the need is for good training
courses (not necessarily certification and all the IMHO nonsense that
comes with that), and a willingness on the part of employers to invest
in upskilling their staff.

You know its funny. Command Prompt, OTG-Inc, SRA and Big Nerd Ranch
*all* offer training.

Last I checked, OTG had to cancel classes because of lack of demand
(please verify Chander).

Well that is interesting, so maybe there is no need for certification
yet? or do you think employers are wanting folks that someone *else* has
trained (or certified).

Cheers

Mark

#41

jd@commandprompt.com

almost 19 years ago

In reply to: Mark Kirkwood (#40)

Re: About PostgreSQL certification

Mark Kirkwood wrote:

Joshua D. Drake wrote:

To cure the shortage of experienced Postgres folks there is only one
solution - err, more experience! So the need is for good training
courses (not necessarily certification and all the IMHO nonsense that
comes with that), and a willingness on the part of employers to invest
in upskilling their staff.

You know its funny. Command Prompt, OTG-Inc, SRA and Big Nerd Ranch
*all* offer training.

Last I checked, OTG had to cancel classes because of lack of demand
(please verify Chander).

Well that is interesting, so maybe there is no need for certification
yet? or do you think employers are wanting folks that someone *else* has
trained (or certified).

I believe that there is a market for Training, certainly (just watch
CMDs website in the next couple of weeks). However I believe that the
market is specific and nitch.

As far as certification, and I guarantee you Theo's experience is
different, I believe certification is dying. Certification used to make
sense, computing was relatively new tech. Keep in mind that the common
user base for computing is only 12-15 years old.

10 years ago.. you literally didn't know if the guy you were hiring new
his stuff or was lying through his teeth. You were a manager who grew up
with green ledger on wide print dot matrix. That little paper said,
"This guy has at least read the book".

Today? Its different in most markets. Theo and I for the most part don't
share market which is why I think his experience is different. My market
is say the 90% market, that is to say that I focus on a more general
service of PostgreSQL.

The people I deal with are FOSS people or people looking pointedly and
moving to FOSS. They don't give a winkle, dinkle about certification.
Most of my customers are business tech savvy, meaning they know outlook,
the know word, they understand the web and the internet.

They are not programmers but the know the difference between:

We are going to create a synergetic alliance of vertical technologies to
integrate your diverse infrastructure.

and

We are going to install Samba so windows, linux and apple can all talk
to a single file share. It will take a day.

Then again, I haven't had to write a resume in 10 years, what the hell
do I know.

Sincerely,

Joshua D. Drake

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#42

John Bartlett

johnb@fast.fujitsu.com.au

almost 19 years ago

In reply to: Richard Troy (#25)

Re: Updateable cursors

Hi Richard,

Thanks for your comments.

I can see where you are coming from but I am not sure if a new log entry
would be such a good idea. The result of creating such a low level log could
be to increase the amount of logging by a rather large amount.

However, the system catalogue will contain an entry that enables a cursor to
be identified as updatable.

Regards,
John Bartlett
Software Development Engineer
Fujitsu Australia Software Technology
14 Rodborough Road, Frenchs Forest NSW 2086
Tel: +61 2 9452 9161
Fax: +61 2 9975 2899
Email: johnb@fast.fujitsu.com.au
Web site: www.fastware.com

-----Original Message-----
From: pgsql-hackers-owner@postgresql.org
[mailto:pgsql-hackers-owner@postgresql.org] On Behalf Of Richard Troy
Sent: Wednesday, 24 January 2007 4:37 AM
To: FAST PostgreSQL
Cc: PostgreSQL-development
Subject: Re: [HACKERS] Updateable cursors

On Wed, 24 Jan 2007, FAST PostgreSQL wrote:

We are trying to develop the updateable cursors functionality into
Postgresql. I have given below details of the design and also issues we

are

facing. Looking forward to the advice on how to proceed with these

issues.

Rgds,
Arul Shaji

Hi Arul,

...I can see people are picking apart the implementation details so you're
getting good feedback on your ambitious proposal. Looks like you've put a
lot of thought/work into it.

Regards,
Richard

--
Richard Troy, Chief Scientist
Science Tools Corporation
510-924-1363 or 202-747-1263
rtroy@ScienceTools.com, http://ScienceTools.com/

---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql.org so that your
message can get through to the mailing list cleanly

If you do not wish to receive commercial email messages from Fujitsu Australia Software Technology Pty Ltd, please email unsubscribe@fast.fujitsu.com.au

#43

John Bartlett

johnb@fast.fujitsu.com.au

almost 19 years ago

In reply to: Simon Riggs (#14)

Re: Updateable cursors

Hi Simon,

Thanks for your comments.

The reason for those 5 options is to consider different means to cover the
Prepared Stmt requirement where the different stages of processing are
actually in different transactions.

-----Original Message-----
From: pgsql-hackers-owner@postgresql.org
[mailto:pgsql-hackers-owner@postgresql.org] On Behalf Of Simon Riggs
Sent: Tuesday, 23 January 2007 11:12 PM
To: FAST PostgreSQL
Cc: PostgreSQL-development
Subject: Re: [HACKERS] Updateable cursors

On Wed, 2007-01-24 at 02:42 +1100, FAST PostgreSQL wrote:

In the UPDATE or DELETE statements the 'WHERE CURRENT OF <cursor_name>'
clause results in the cursor name being placed in the UpdateStmt or
DeleteStmt structure. During the processing of the functions -
transformDeleteStmt() and transformUpdateStmt() - the cursor name is used

obtain a pointer to the related Portal structure

To support prepared statements we'd need to do this name lookup just
once, so that the Update/Delete stmt can record which Portal to look at
for the current tuple.

and the tuple affected by
the current UPDATE or DELETE statement is extracted from the Portal, where

has been placed as the result of a previous FETCH request. At this point

all

the information for the UPDATE or DELETE statement is available so the
statements can be transformed into standard UPDATE or DELETE statements

and

sent for re-write/planning/execution as usual.

2.5 Changes to the Executor
-------------------------------
There are various options that have been considered for this part of the
enhancement. These are described in the sections below.

Option 1 MVCC Via Continuous Searching of Database

The Executor is to be changed in the following ways:
1) When the FETCH statement is executed the id of the resulting tuple

extracted and passed back to the Portal structure to be saved to indicate

the

cursor is currently positioned on a tuple.
2) When the UPDATE or DELETE request is executed the tuple id

previously

FETCHed is held in the QueryDesc structure so that it can be compared with

the tuple ids returned from the TidScan node processed prior to the actual

UPDATE / DELETE node in the plan. This enables a decision to be made as to

whether the tuple held in the cursor is visible to the UPDATE / DELETE
request according to the rules of concurrency. The result is that, at the
cost of repeatedly searching the database at each UPDATE / DELETE command,

the hash table is no longer required.
This approach has the advantage that there is no hash table held in memory

on disk so it will not be memory intensive but will be processing

intensive.

Do you have a specific example that would cause problems? It's much
easier to give examples that might cause problems and discuss those.

So I'm not sure why there's a big need for any of the 5 options, yet.

--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com

---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend

If you do not wish to receive commercial email messages from Fujitsu Australia Software Technology Pty Ltd, please email unsubscribe@fast.fujitsu.com.au

#44

Jim Nasby

decibel@decibel.org

almost 19 years ago

In reply to: Joshua D. Drake (#39)

Re: About PostgreSQL certification

On Jan 23, 2007, at 5:50 PM, Joshua D. Drake wrote:

To cure the shortage of experienced Postgres folks there is only one
solution - err, more experience! So the need is for good training
courses (not necessarily certification and all the IMHO nonsense that
comes with that), and a willingness on the part of employers to
invest
in upskilling their staff.

You know its funny. Command Prompt, OTG-Inc, SRA and Big Nerd Ranch
*all* offer training.

Last I checked, OTG had to cancel classes because of lack of demand
(please verify Chander).

That may be OTG's problem then... I've personally taught a number of
classes since I started with EnterpriseDB (PostgreSQL classes, not
EnterpriseDB ones). Granted, this training is for existing customers,
but I believe it speaks to the demand that's out there.

And while certification might not mean much to people knowledgeable
enough to tell if someone has clue, I suspect that as PostgreSQL
grows in popularity more people will look at training (especially for
people that don't have "PostgreSQL" stamped all over their resume).

On the other hand, any time I find someone interesting in pushing
their career towards PostgreSQL I always tell them the same thing:
get on the mailing list and start helping folks. Perhaps that's
ultimately all the certification we'll ever need.
--
Jim Nasby jim@nasby.net
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)

#45

Zeugswetter Andreas ADI SD

simon@2ndquadrant.com

almost 19 years ago

In reply to: John Bartlett (#43)

Re: Updateable cursors

On Wed, 2007-01-24 at 14:54 +1100, John Bartlett wrote:

The reason for those 5 options is to consider different means to cover the
Prepared Stmt requirement where the different stages of processing are
actually in different transactions.

John,

Thanks for explaining.

Wow! I've never come across such a requirement before, personally and
hadn't even imagined anybody would want to do this.

ISTM the main use for positioned UPDATE/DELETE is for a single
transaction to first open a cursor and then loop around doing FETCH and
then positioned UPDATE/DELETE on that cursor.

It would make the implementation considerably easier to limit the
initial implementation to only work using WITHOUT HOLD cursors (the
default). This will allow you to cache the ctid, rather than re-seeking
via the index, so will offer considerably better performance also.

That is also the safe thing to do, since PostgreSQL's implementation of
WITH HOLD cursors doesn't leave the rows locked. That can lead to the
rows being deleted from under the cursor, for which the standard is
unclear as to whether that is acceptable, or not.

AFAICS the SQL Standard also requires that the positioned Update/Delete
also effect only a single row. When using WITH HOLD cursors the desired
row's ctid may have changed. Re-executing the original WHERE condition
might easily reveal more than one row where previously there was only
one. The cursor itself provides no mechanism for telling rows apart in
that circumstance when no Primary Key is defined on the table. We can
surround that with various checks, maybe. ISTM that even allowing this
using WITH HOLD cursors seems likely to be both a poor-performing and
fragile application programming technique.

I'd suggest we add the combination of WITH HOLD cursors and positioned
updates to the small pile of SQL standard items we don't really want to
support for practical reasons.

At very least, I'd suggest we do the straightforward part of this for
8.3 and see whether we want a more full implementation in later
releases.

--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com

#46

ZeugswetterA@spardat.at

almost 19 years ago

In reply to: Simon Riggs (#45)

Re: Updateable cursors

That is also the safe thing to do, since PostgreSQL's implementation

WITH HOLD cursors doesn't leave the rows locked. That can lead to the
rows being deleted from under the cursor, for which the standard is
unclear as to whether that is acceptable, or not.

Um, the default use case is to "intent exclusive" lock the current row,
so you can do some calculations on columns inside the application
without
them changing in the meantime.
So, imho that lock is a substantial feature of FOR UPDATE cursors.
The lock is usually freed as soon as you fetch the next row.
In MVCC db's it is also a method to read a guaranteed up to date
version.

Andreas

#47

John Zubac

jzubac@internetsecure.com

almost 19 years ago

In reply to: Jim Nasby (#44)

Re: About PostgreSQL certification

Hi Theo

I find your statements about Postgres being a huge business risk pretty laughable. First of all, Postgres is based on SQL92 and SQL99 standards which means that most scripts are pretty much the same compared to MSSQL and Oracle. The only thing I have seen to learn are the postgres datatypes. Big deal! PGAdmin III will write most scripts for you and that too is pretty much free. I dealt with it when we started learning and using postgres. I only had experience in Oracle and MSSQL.

Also comparing Postgres to MYSQL is also pretty funny, since there are instances of MYSQL LOSING databases due to corruption because they do not have PITR and their transaction rollback feature did not work properly last time I checked. This is really a issue of people being close minded to great database software and not being able to sell it to their superiors.

This is the way I sold postgres to my boss. It is opensource (low cost), all the features of MSSQL and then some, WAY FASTER than MSSQL on a BSD platform, very good recovery when the database gets corrupted (this happens to all databases from user error usually), and lastly you can always migrate the data to another database if you don't like postgres in the end.

John Zubac

-----Original Message-----
From: pgsql-hackers-owner@postgresql.org
[mailto:pgsql-hackers-owner@postgresql.org]On Behalf Of Theo
Schlossnagle
Sent: Tuesday, January 23, 2007 5:20 PM
To: Joshua D. Drake
Cc: Theo Schlossnagle; Mark Kirkwood; David Fetter; Iannsp;
PostgreSQL-development
Subject: Re: [HACKERS] About PostgreSQL certification

On Jan 23, 2007, at 5:14 PM, Joshua D. Drake wrote:

Get a CCIE and tell me that again :-) When you are handed a
complicated
network of routers and switches running all sorts of version of
IOS and
CatOS and you go to lunch, they break it and you have a certain time
allotment to fix it all.

Most certifications are not simple multiple choice quizes. Just the
ones you hear about -- the ones that suck.

I think seeing relevant training courses + experience on a CV trumps
certification anytime - unfortunately a lot of folks out there are
mesmerized by shiny certificates....

Sure. But experience is very hard to get. And since people with
PostgreSQL experience are limited, companies adopting it need a good
second option -- certified people.

They aren't limited, just all employed ;)

--
Theo

// Theo Schlossnagle
// CTO -- http://www.omniti.com/~jesus/
// OmniTI Computer Consulting, Inc. -- http://www.omniti.com/

Import Notes

Resolved by subject fallback

#48

simon@2ndquadrant.com

almost 19 years ago

In reply to: Zeugswetter Andreas ADI SD (#46)

Re: Updateable cursors

On Wed, 2007-01-24 at 14:27 +0100, Zeugswetter Andreas ADI SD wrote:

That is also the safe thing to do, since PostgreSQL's implementation

of

WITH HOLD cursors doesn't leave the rows locked. That can lead to the
rows being deleted from under the cursor, for which the standard is
unclear as to whether that is acceptable, or not.

Um, the default use case is to "intent exclusive" lock the current row,
so you can do some calculations on columns inside the application
without
them changing in the meantime.
So, imho that lock is a substantial feature of FOR UPDATE cursors.
The lock is usually freed as soon as you fetch the next row.
In MVCC db's it is also a method to read a guaranteed up to date
version.

Completely agree.

The standard doesn't say it, but it might be taken to imply that locks
continue to be held, as with 2PC, and released when the cursor is
closed. But I'm not really sure I'd want that either, IMHO.

--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com

#49

jesus@omniti.com

almost 19 years ago

In reply to: John Zubac (#47)

Re: About PostgreSQL certification

On Jan 24, 2007, at 9:50 AM, John Zubac wrote:

I find your statements about Postgres being a huge business risk
pretty laughable. First of all, Postgres is based on SQL92 and
SQL99 standards which means that most scripts are pretty much the
same compared to MSSQL and Oracle. The only thing I have seen to
learn are the postgres datatypes. Big deal! PGAdmin III will write
most scripts for you and that too is pretty much free. I dealt with
it when we started learning and using postgres. I only had
experience in Oracle and MSSQL.

If that's the only thing you had to learn, we aren't talking about
the same risks. Datatypes are developer level differences. Tuning,
sizing, disaster recovery planning, backups, differences in or lack
of enterprise features and integration -- these are all very
different between databases and fundamental to operating it in an
enterprise environment.

You can laugh if you like. I don't laugh about these things, neither
do our clients. Many have decided to run postgres and that decision
was a good one. Many do not and their decisions were also wise.
Several people at the dayjob, including me, travel and speak on
postgres, database replication, large architecture management, open
source, etc. We promote postgres in many venues. I said: "If you
only want to hire people with extensive experience, you're exposing
yourself to an enormous business risk by adopting postgres."

There simple aren't that many people that have extensive experience.
So you you are hinging the success of your business of one of those
people being available, it _is_ an enormous business risk. My
arguments here are not against postgres, they are for training and
certification -- both help dramatically increase the pool of people
with sufficient experience.

Also comparing Postgres to MYSQL is also pretty funny, since there
are instances of MYSQL LOSING databases due to corruption because
they do not have PITR and their transaction rollback feature did
not work properly last time I checked. This is really a issue of
people being close minded to great database software and not being
able to sell it to their superiors.

It's not funny at all. Just like comparing PostgreSQL to Apache
isn't funny (Covalent did spectacular things legitimizing the use of
Apache in the global 2000). The fact that MySQL has lost data is not
germane to the discussion. There have been bugs in PostgreSQL as
well. And there has been data loss with PostgreSQL and Oracle and
MSSQL. We're talking about business risks due to resource
availability in the job market capable of managing postgresql in an
enterprise environment. And was stating that solid certification
programs can and will increase the availability of those resources
and reduce the risks in adopting postgres as a solution.

I, along with most of the people in the community, believe in
PostgreSQL, believe in the direction development is going in and want
to see adoption increase.

This is the way I sold postgres to my boss. It is opensource (low
cost), all the features of MSSQL and then some, WAY FASTER than
MSSQL on a BSD platform, very good recovery when the database gets
corrupted (this happens to all databases from user error usually),
and lastly you can always migrate the data to another database if
you don't like postgres in the end.

I have no problem representing the positive aspects of postgres. I
am also not blind to its shortcomings. We manage one of the larger
postgres instances out there -- I know its pros and cons well.

// Theo Schlossnagle
// CTO -- http://www.omniti.com/~jesus/
// OmniTI Computer Consulting, Inc. -- http://www.omniti.com/

#50

http://www.forbes.com/lists/2006/18/06f2000_The-Forbes-2000_Rank.html

jd@commandprompt.com

almost 19 years ago

In reply to: Theo Schlossnagle (#49)

Re: About PostgreSQL certification

Also comparing Postgres to MYSQL is also pretty funny, since there are
instances of MYSQL LOSING databases due to corruption because they do
not have PITR and their transaction rollback feature did not work
properly last time I checked. This is really a issue of people being
close minded to great database software and not being able to sell it
to their superiors.

It's not funny at all. Just like comparing PostgreSQL to Apache isn't
funny (Covalent did spectacular things legitimizing the use of Apache in
the global 2000). The fact that MySQL has lost data is not germane to

This is the point of this thread that I think people are severely missing.

(Covalent did spectacular things legitimizing the use of Apache in the
global 2000)

It is also about my point that Theo and I share different markets. In
Theo's world his arguments are 100% correct, imo.

I would garner that less than 1% of the PostgreSQL experts out there can
speak to the global 2000 requirements. The global 2000 includes people
like GM, Wal-Mart and Sony.

These organizations have diverse and extreme requirements that only some
of us have ever even been exposed to.

Case in point, one of my customers recently spoke to me about moving a
critical system to PostgreSQL. This system, if down will cost the
customer several million (that is 7 digits) an hour.

How many on this thread can honestly say that they have a clue what type
of business volume that is?

Sincerely,

Johsua D. Drake

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

#51

rick@alpinenetworking.com

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: 10 weeks to feature freeze (Pending Work)

I thought that the following todo item just barely missed 8.2:

"Allow a warm standby system to also allow read-only statements [pitr]
This is useful for checking PITR recovery."

I assume it's not on this list either because it is already complete and
slated for 8.3, or it is going to take too long to make it into 8.3 or
it has been rejected as a good idea entirely or it's just not big enough
of a priority for anyone to push for it to get into 8.3.

It is the one feature that would make the most difference to me as it
would allow me to very easily set up a server for reporting purposes
that could always be within minutes of the live data. I know there are
other solutions for this but if this feature is just around the corner
it would be my first choice.

Does anyone know the status of this feature?

Thanks,

Rick Gigger

Joshua D. Drake wrote:

Show quoted text

Or so... :)

Thought I would do a poll of what is happening in the world for 8.3. I have:

Alvaro Herrera: Autovacuum improvements (maintenance window etc..)
Gavin Sherry: Bitmap Indexes (on disk), possible basic Window functions
Jonah Harris: WITH/Recursive Queries?
Andrei Kovalesvki: Some Win32 work with Magnus
Magnus Hagander: VC++ support (thank goodness)
Heikki Linnakangas: Working on Vacuum for Bitmap Indexes?
Oleg Bartunov: Tsearch2 in core
Neil Conway: Patch Review (including enums), pg_fcache

Vertical projects:

Pavel Stehule: PLpsm
Alexey Klyukin: PLphp
Andrei Kovalesvki: ODBCng

I am sure there are more, the ones with question marks are unknowns but
heard of in the ether somewhere. Any additions or confirmations?

Sincerely,

Joshua D. Drake

#52

Tom Lane

tgl@sss.pgh.pa.us

almost 19 years ago

In reply to: Rick Gigger (#51)

Re: 10 weeks to feature freeze (Pending Work)

Rick Gigger <rick@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:
"Allow a warm standby system to also allow read-only statements [pitr]

No, it's a someday-wishlist item; the work involved is not small.

regards, tom lane

#53

Henry B. Hotz

hotz@jpl.nasa.gov

almost 19 years ago

In reply to: Joshua D. Drake (#1)

Re: 10 weeks to feature freeze (Pending Work)

Henry B. Hotz: GSSAPI authentication method for C (FE/BE) and Java
(FE).
Magnus Haglander: SSPI (GSSAPI compatible) authentication method for
C (FE) on Windows.

(That fair Magnus? Or you want to volunteer for BE support as well?)

GSSAPI isn't much more than a functional replacement for Kerberos 5,
but it's supported on lots more platforms. In particular Java and
Windows have native support (as well as Solaris 9).

If anyone is interested I currently have working-but-incomplete
patches to support SASL in C. I've decided not to finish and submit
them because the glue code to make configuration reasonable, and to
allow use of existing Postgres password databases with the password-
based mechanisms is still significant.

On Jan 22, 2007, at 2:16 PM, Joshua D. Drake wrote:

Show quoted text

Or so... :)

Thought I would do a poll of what is happening in the world for
8.3. I have:

Alvaro Herrera: Autovacuum improvements (maintenance window etc..)
Gavin Sherry: Bitmap Indexes (on disk), possible basic Window
functions
Jonah Harris: WITH/Recursive Queries?
Andrei Kovalesvki: Some Win32 work with Magnus
Magnus Hagander: VC++ support (thank goodness)
Heikki Linnakangas: Working on Vacuum for Bitmap Indexes?
Oleg Bartunov: Tsearch2 in core
Neil Conway: Patch Review (including enums), pg_fcache

Vertical projects:

Pavel Stehule: PLpsm
Alexey Klyukin: PLphp
Andrei Kovalesvki: ODBCng

I am sure there are more, the ones with question marks are unknowns
but
heard of in the ether somewhere. Any additions or confirmations?

Sincerely,

Joshua D. Drake

--

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/
donate
PostgreSQL Replication: http://www.commandprompt.com/products/

---------------------------(end of
broadcast)---------------------------
TIP 5: don't forget to increase your free space map settings

#54

Gregory Stark

stark@enterprisedb.com

almost 19 years ago

In reply to: Rick Gigger (#51)

Re: 10 weeks to feature freeze (Pending Work)

"Rick Gigger" <rick@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:

"Allow a warm standby system to also allow read-only statements [pitr]
This is useful for checking PITR recovery."

No, nobody worked on it prior to 8.2. Afaik there's still nobody working on
it. It's not trivial. Consider for example that your read-only query would
still need to come up with a snapshot and there's nowhere currently to find
out what transactions were in-progress at that point in the log replay.

There's also the problem that currently WAL replay doesn't take have allow for
any locking so there's no way for read-only queries to protect themselves
against the WAL replay thrashing the buffer pages they're looking at.

It does seem to be doable and I agree it would be a great feature, but as far
as I know nobody's working on it for 8.3.

--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com

#55

Gregory Stark

stark@enterprisedb.com

almost 19 years ago

In reply to: Rick Gigger (#51)

Re: 10 weeks to feature freeze (Pending Work)

"Rick Gigger" <rick@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:

"Allow a warm standby system to also allow read-only statements [pitr]
This is useful for checking PITR recovery."

It does seem to be doable and I agree it would be a great feature, but as far
as I know nobody's working on it for 8.3.

--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com

#56

Stephen Frost

sfrost@snowman.net

almost 19 years ago

In reply to: Henry B. Hotz (#53)

Re: 10 weeks to feature freeze (Pending Work)

* Henry B. Hotz (hotz@jpl.nasa.gov) wrote:

If anyone is interested I currently have working-but-incomplete
patches to support SASL in C. I've decided not to finish and submit
them because the glue code to make configuration reasonable, and to
allow use of existing Postgres password databases with the password-
based mechanisms is still significant.

I'd certainly like to take a look at it. I'm not entirely sure I follow
what you mean by 'allow use of existing Postgres password databases'-
I'm not sure SASL support requires that ability (after all, if they want
to use the 'md5' or similar mechanism they can with the current
protocol). Or am I missing something about how the SASL implementation
is done or intended to be used? I'd tend to think it'd mainly be used
as a mechanism to support other authentication mechanisms which don't
use the internal Postgres passwords...

Thanks,

Stephen

#57

Henry B. Hotz

hotz@jpl.nasa.gov

almost 19 years ago

In reply to: Stephen Frost (#56)

Re: 10 weeks to feature freeze (Pending Work)

Sent directly. Anyone else who's interested can have a copy. Just
email me.

I *think* it's structurally sound. Please tell me if you find a
problem. It lacks a lot: proper specification of required security
properties, a way to specify different mechanism lists for local,
vice TCP, vice SSL connections, authN name to authZ name mapping,
most seriously I didn't implement security layers. Lots of debug
checking still needed.

OTOH it works on MacOS 10.4 G4 client and Intel server.

As to the Postgres password database: If you use the DIGEST-MD5
mechanism, then you could get a secure, encrypted connection with no
setup except the PG password. Also it would have made it easier for
people to migrate from the current stuff to SASL.

SASL *could* do everything that *any* of the current auth methods can
do (OK, except ident) and then some. I thought that exporting all
that code and functionality to a standard library would be a good
thing in the long run. The down side is that completely replacing
the existing framework would require SASL libraries readily available
on *all* platforms that PG supports, and Windows doesn't. The
Windows SASL API's turn out to be only available on 2K3 server, and
have never been publicly tested for interoperability with the
standard Unix library.

I still believe in SASL. I know the Cyrus SASL library has become
pretty ubiquitous on Unix platforms. I wish there were a simpler C
API than Cyrus. Java 1.4.2 and up supports it. There are ways it
could be provided on Windows, but not within the level of effort that
Magnus or I can devote to the problem.

---------

For GSSAPI, there is published interop code for the Windows SSPI at
<http://web.mit.edu/jaltman/Public/kfw/gss/>. It's more places than
SASL is. Down side is it doesn't do much that the current Krb5 code
doesn't do.

Structurally the GSSAPI mods will be very similar to the SASL ones I
already did.

On Jan 26, 2007, at 7:16 PM, Stephen Frost wrote:

* Henry B. Hotz (hotz@jpl.nasa.gov) wrote:

If anyone is interested I currently have working-but-incomplete
patches to support SASL in C. I've decided not to finish and submit
them because the glue code to make configuration reasonable, and to
allow use of existing Postgres password databases with the password-
based mechanisms is still significant.

I'd certainly like to take a look at it. I'm not entirely sure I
follow
what you mean by 'allow use of existing Postgres password databases'-
I'm not sure SASL support requires that ability (after all, if they
want
to use the 'md5' or similar mechanism they can with the current
protocol). Or am I missing something about how the SASL
implementation
is done or intended to be used? I'd tend to think it'd mainly be used
as a mechanism to support other authentication mechanisms which don't
use the internal Postgres passwords...

Thanks,

Stephen

------------------------------------------------------------------------
The opinions expressed in this message are mine,
not those of Caltech, JPL, NASA, or the US Government.
Henry.B.Hotz@jpl.nasa.gov, or hbhotz@oxy.edu

#58

Richard Troy

rtroy@ScienceTools.com

almost 19 years ago

In reply to: John Bartlett (#42)

Re: Updateable cursors

On Wed, 24 Jan 2007, John Bartlett wrote:
[regarding optional DBA/SysAdmin logging of Updateable Cursors]

I can see where you are coming from but I am not sure if a new log entry
would be such a good idea. The result of creating such a low level log could
be to increase the amount of logging by a rather large amount.

Given that logging can be controlled via the contents of postgresql.conf,
this sounds like an answer from someone who's never had to support a
production environment; Putting a check for log_min_error_statement being
set to, say, info, hardly seems like a big burden to me. A casual study of
the controls in postgresql.conf reveals we already have many controlls to
get things logged when we want/need them - all of which were deemed
appropriate previously. So ISTM that if the DBA/SysAdmin thinks they need
the information, who are you to tell them, in effect, "No, I don't want
you to have to spend any of your machine's performace giving you the
information you need?"

Help your user by giving them information when they want it. ... Do you
argue that this is useless information?

Richard

--
Richard Troy, Chief Scientist
Science Tools Corporation
510-924-1363 or 202-747-1263
rtroy@ScienceTools.com, http://ScienceTools.com/

#59

Magnus Hagander

magnus@hagander.net

almost 19 years ago

In reply to: Henry B. Hotz (#53)

Re: 10 weeks to feature freeze (Pending Work)

Henry B. Hotz wrote:

Henry B. Hotz: GSSAPI authentication method for C (FE/BE) and Java (FE).
Magnus Haglander: SSPI (GSSAPI compatible) authentication method for C
(FE) on Windows.

(That fair Magnus? Or you want to volunteer for BE support as well?)

Seems fair and about what we discussed. And no, I won't volunteer as
long as you're on it - not sure I'll have the time to do it all in time.

GSSAPI isn't much more than a functional replacement for Kerberos 5, but
it's supported on lots more platforms. In particular Java and Windows
have native support (as well as Solaris 9).

Yeah, getting rid of the dependency on MIT KRB5 on windows would be very
nice indeeed.

//Magnus

#60

Henry B. Hotz

hotz@jpl.nasa.gov

almost 19 years ago

In reply to: Magnus Hagander (#59)

Re: 10 weeks to feature freeze (Pending Work)

On Jan 29, 2007, at 9:49 AM, Magnus Hagander wrote:

Henry B. Hotz wrote:

Henry B. Hotz: GSSAPI authentication method for C (FE/BE) and
Java (FE).
Magnus Haglander: SSPI (GSSAPI compatible) authentication method
for C
(FE) on Windows.

(That fair Magnus? Or you want to volunteer for BE support as well?)

Seems fair and about what we discussed. And no, I won't volunteer as
long as you're on it - not sure I'll have the time to do it all in
time.

I'm only volunteering BE for Unix, not Windows. Not sure we need BE
for Windows for 8.3 though. This is enough.
------------------------------------------------------------------------
The opinions expressed in this message are mine,
not those of Caltech, JPL, NASA, or the US Government.
Henry.B.Hotz@jpl.nasa.gov, or hbhotz@oxy.edu

#61

Magnus Hagander

magnus@hagander.net

almost 19 years ago

In reply to: Henry B. Hotz (#60)

Re: 10 weeks to feature freeze (Pending Work)

On Mon, Jan 29, 2007 at 12:44:51PM -0800, Henry B. Hotz wrote:

On Jan 29, 2007, at 9:49 AM, Magnus Hagander wrote:

Henry B. Hotz wrote:

Henry B. Hotz: GSSAPI authentication method for C (FE/BE) and
Java (FE).
Magnus Haglander: SSPI (GSSAPI compatible) authentication method
for C
(FE) on Windows.

(That fair Magnus? Or you want to volunteer for BE support as well?)

Seems fair and about what we discussed. And no, I won't volunteer as
long as you're on it - not sure I'll have the time to do it all in
time.

I'm only volunteering BE for Unix, not Windows. Not sure we need BE
for Windows for 8.3 though. This is enough.

Oh certainly, I'm thinking BE on windows as well, but not sure if we'll
have it for 8.3. We need to have frontend, so we have the same support
as we have for krb5. Backend is a bonus, but it'd be nice to have it.

//Magnus

#62

Chander Ganesan

chander@otg-nc.com

almost 19 years ago

In reply to: Joshua D. Drake (#39)

Re: About PostgreSQL certification

Joshua D. Drake wrote:

To cure the shortage of experienced Postgres folks there is only one
solution - err, more experience! So the need is for good training
courses (not necessarily certification and all the IMHO nonsense that
comes with that), and a willingness on the part of employers to invest
in upskilling their staff.

You know its funny. Command Prompt, OTG-Inc, SRA and Big Nerd Ranch
*all* offer training.

Last I checked, OTG had to cancel classes because of lack of demand
(please verify Chander).

We tried to offer classes in Santa Clara, CA (and may do so again), but
didn't have sufficient demand to run them. There were 2 people that had
expressed interest in that class. However, I will say that in 2006 that
was the only class we canceled (keep in mind though, that we run courses
at our headquarters *regardless* of the number of students enrolled -
it's our policy to not cancel classes here...)

On another note, I will say that we're doing well enough to support the
project through SPI...

Command Prompt currently only trains corps with 12+ people per class so
we are a bit different.

Sinerely,

Joshua D. Drake

--
Chander Ganesan
The Open Technology Group
One Copley Parkway, Suite 210
Morrisville, NC 27560
Phone: 877-258-8987/919-463-0999
http://www.otg-nc.com

#63

Chander Ganesan

chander@otg-nc.com

almost 19 years ago

In reply to: Iannsp (#16)

Re: About PostgreSQL certification

Ivo,

Iannsp wrote:

Hello,
I did like to know what you think about the postgresql certifications
provided for

PostgreSQL CE
http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst
http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the clients.
Make difference to be certified?

IMHO, the SRA certification has been around for awhile, and I believe it
has some credibility in Japan...While I'm not sure what its credibility
is like here in the US - the fact that it has credibility in Japan is a
big plus .

The CertFirst certification (examsonline.com), seems to be administered
online (as opposed to SRA's which is at a PearsonVUE test center) -
which basically means that it's open book, open note, call your friend,
copy the questions, etc. It also seems that CertFirst runs the
certification themselves under what appears to be a shell company called
"examsonline". It looks to be more of a marketing ploy than anything
else....

Based on the fact that they are a provider of training under the WIA act
in IL, I'd suspect that they need a certification so that they can sell
their programs to the unemployed folks that are getting free training on
the gov'ts dime.

thanks for advanced.

Ivo Nascimento.

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster

--
Chander Ganesan
The Open Technology Group
One Copley Parkway, Suite 210
Morrisville, NC 27560
Phone: 877-258-8987/919-463-0999
http://www.otg-nc.com

#64

Chander Ganesan

chander@otg-nc.com

almost 19 years ago

In reply to: Chander Ganesan (#63)

Re: About PostgreSQL certification

Chander Ganesan wrote:

Ivo,

Iannsp wrote:

Hello,
I did like to know what you think about the postgresql certifications
provided for

PostgreSQL CE
http://www.sraoss.co.jp/postgresql-ce/news_en.html

CertFirst
http://www.certfirst.com/postgreSql.htm

My question is about the validate of this certification for the clients.
Make difference to be certified?

IMHO, the SRA certification has been around for awhile, and I believe
it has some credibility in Japan...While I'm not sure what its
credibility is like here in the US - the fact that it has credibility
in Japan is a big plus .

The CertFirst certification (examsonline.com), seems to be
administered online (as opposed to SRA's which is at a PearsonVUE test
center) - which basically means that it's open book, open note, call
your friend, copy the questions, etc. It also seems that CertFirst
runs the certification themselves under what appears to be a shell
company called "examsonline". It looks to be more of a marketing ploy
than anything else....

Correction...I just checked and it looks like they've updated their web
site and no longer refer to the examsonline online exam...so I'm not
sure where/what their exam entails now. Their site used to refer to an
exam through examsonline.com ... You'll have to contact them for details...

Based on the fact that they are a provider of training under the WIA
act in IL, I'd suspect that they need a certification so that they can
sell their programs to the unemployed folks that are getting free
training on the gov'ts dime.

thanks for advanced.

Ivo Nascimento.

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster

--
Chander Ganesan
The Open Technology Group
One Copley Parkway, Suite 210
Morrisville, NC 27560
Phone: 877-258-8987/919-463-0999
http://www.otg-nc.com

#65

Andrew Hammond

andrew.george.hammond@gmail.com

almost 19 years ago

In reply to: Tom Lane (#52)

Re: 10 weeks to feature freeze (Pending Work)

On Jan 26, 2:38 pm, t...@sss.pgh.pa.us (Tom Lane) wrote:

Rick Gigger <r...@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:
"Allow a warm standby system to also allow read-only statements [pitr]

No, it's a someday-wishlist item; the work involved is not small.

Slony1 has supported log-shipping replication for about a year now. It
provides similar functionality.

Andrew

#66

rick@alpinenetworking.com

almost 19 years ago

In reply to: Tom Lane (#52)

Re: 10 weeks to feature freeze (Pending Work)

Tom Lane wrote:

Rick Gigger <rick@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:
"Allow a warm standby system to also allow read-only statements [pitr]

No, it's a someday-wishlist item; the work involved is not small.

Thanks,very much for the info. I'm not sure why I thought that one was
near completion. I can now come up with an alternative plan.

#67

rick@alpinenetworking.com

almost 19 years ago

In reply to: Andrew Hammond (#65)

Re: 10 weeks to feature freeze (Pending Work)

Andrew Hammond wrote:

On Jan 26, 2:38 pm, t...@sss.pgh.pa.us (Tom Lane) wrote:

Rick Gigger <r...@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:
"Allow a warm standby system to also allow read-only statements [pitr]

No, it's a someday-wishlist item; the work involved is not small.

Slony1 has supported log-shipping replication for about a year now. It
provides similar functionality.

Yes but Slony is much more complicated, has significantly more
administrative overhead, and as far as I can tell is much more likely to
impact my production system than this method would.

Slony is a lot more flexible and powerful but I don't need that. I just
want a backup that is reasonably up to date that I can do queries on and
and failover to in case of hardware failure on my primary db.

I am going to be looking more closely at Slony now that it seems to be
the best option for this. I am not looking forward to how it will
complicate my life though. (Not saying it is bad, just complicated. At
least more complicated than simple postgres log shipping.

#68

rick@alpinenetworking.com

almost 19 years ago

In reply to: Gregory Stark (#55)

Re: 10 weeks to feature freeze (Pending Work)

Gregory Stark wrote:

"Rick Gigger" <rick@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:

"Allow a warm standby system to also allow read-only statements [pitr]
This is useful for checking PITR recovery."

No, nobody worked on it prior to 8.2. Afaik there's still nobody working on
it. It's not trivial. Consider for example that your read-only query would
still need to come up with a snapshot and there's nowhere currently to find
out what transactions were in-progress at that point in the log replay.

There's also the problem that currently WAL replay doesn't take have allow for
any locking so there's no way for read-only queries to protect themselves
against the WAL replay thrashing the buffer pages they're looking at.

It does seem to be doable and I agree it would be a great feature, but as far
as I know nobody's working on it for 8.3.

Thanks again for the update.

#69

Jim Nasby

decibel@decibel.org

almost 19 years ago

In reply to: Andrew Hammond (#65)

Re: 10 weeks to feature freeze (Pending Work)

On Feb 5, 2007, at 12:53 PM, Andrew Hammond wrote:

On Jan 26, 2:38 pm, t...@sss.pgh.pa.us (Tom Lane) wrote:

Rick Gigger <r...@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:
"Allow a warm standby system to also allow read-only statements
[pitr]

No, it's a someday-wishlist item; the work involved is not small.

Slony1 has supported log-shipping replication for about a year now. It
provides similar functionality.

Not really....

1) It's not possible for a PITR 'slave' to fall behind to a state
where it will never catch up, unless it's just on inadequate
hardware. Same isn't true with slony.
2) PITR handles DDL seamlessly
3) PITR is *much* simpler to configure and maintain
--
Jim Nasby jim@nasby.net
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)

#70

Andrew Hammond

andrew.george.hammond@gmail.com

almost 19 years ago

In reply to: Jim Nasby (#69)

Re: 10 weeks to feature freeze (Pending Work)

On 2/6/07, Jim Nasby <decibel@decibel.org> wrote:

On Feb 5, 2007, at 12:53 PM, Andrew Hammond wrote:

On Jan 26, 2:38 pm, t...@sss.pgh.pa.us (Tom Lane) wrote:

Rick Gigger <r...@alpinenetworking.com> writes:

I thought that the following todo item just barely missed 8.2:
"Allow a warm standby system to also allow read-only statements
[pitr]

No, it's a someday-wishlist item; the work involved is not small.

Slony1 has supported log-shipping replication for about a year now. It
provides similar functionality.

Not really....

1) It's not possible for a PITR 'slave' to fall behind to a state
where it will never catch up, unless it's just on inadequate
hardware. Same isn't true with slony.

I imagine that there are ways to screw up WAL shipping too, but there
are plenty more ways to mess up slony.

2) PITR handles DDL seamlessly
3) PITR is *much* simpler to configure and maintain

4) You need 3 databases to do log shipping using slony1. An origin, a
subscriber which generates the logs and obviously the log-replica.

All of which is why I qualified my statement with "similar".

#71