vacuumdb --jobs deadlock: how to avoid pg_catalog ?

On 5/14/20 8:35 AM, Eduard Rozenberg wrote:

Hello there,

I'm a long-time postgres user but vacuumdb newbie :).

Using vacuumdb (v12.2) with '---jobs' to vacuum a v9.6 database on
localhost with parallel processes:

$ time vacuumdb --full --echo -U postgres --jobs=8 -d mydatabase_test -p
5434 --password

As shown below I ran into the (100%?) predictable/expected deadlock with
sys catalog tables documented in the vacuumdb doc page
(https://www.postgresql.org/docs/12/app-vacuumdb.html).

Question: I specified "-d mydatabase_test" - why is it also doing a full
vac on the system catalog (postgres db) which is practically guaranteed

The postgres database and the system catalogs are two separate things.

to cause a deadlock? I don't need or want it to do a full vac on the
postgres db as I can do that myself later if I need to, and it takes no
time compared to full vac of my own 7 TB database that I am "extremely
eager" to parallelize.

See:

https://www.postgresql.org/docs/12/sql-vacuum.html

"FULL

Selects �full� vacuum, which can reclaim more space, but takes much
longer and exclusively locks the table. This method also requires extra
disk space, since it writes a new copy of the table and doesn't release
the old copy until the operation is complete. Usually this should only
be used when a significant amount of space needs to be reclaimed from
within the table.
"

Pretty sure all that rewriting of tables bloats the system catalogs and
so they get a VACUUM FULL.

The question here is why is it necessary to do a VACUUM FULL in the
first place?

Is my only option list all schemas.tables and feed to vacuumdb as "-t
schema1.table1 -t schema1.table2 ........" ? That's manageable but
unfortunate as it creates more work (having to list all tables) and adds
additional work for automation scripts as well.

I've tried searching forums/goog for past issues related to this. I did
find an old discussion about possible enhancements to "vacuumdb --jobs"
to help the user avoid these predictable deadlocks, where the conclusion
was to let the user figure out a way to deal with it such as via "-t
table -t table -t table .......".

Maybe "--jobs" should only apply to the user's own databases, and when
vacuumdb is working on system catalog tables it should apply a strictly
serial/sequential approach to handle those tables?

Thanks!

------------------------------------------
output from vacuumdb's --echo
------------------------------------------
...
VACUUM (FULL) mycompanytesting.sometable1;
VACUUM (FULL) index.mydatabasestd_missing;
VACUUM (FULL) mycompanytesting.sometable2;
VACUUM (FULL) mycompanytesting.sometable3;
VACUUM (FULL) pg_catalog.pg_attribute;
VACUUM (FULL) mycompanydata.sometable4;
VACUUM (FULL) pg_catalog.pg_statistic;
VACUUM (FULL) mycompanytesting.sometable5;
VACUUM (FULL) pg_catalog.pg_proc;
VACUUM (FULL) pg_catalog.pg_depend;
VACUUM (FULL) pg_catalog.pg_class;
VACUUM (FULL) pg_catalog.pg_type;
vacuumdb: error: vacuuming of database "mydatabase_test" failed: ERROR:
�deadlock detected
DETAIL: �Process 28183 waits for AccessShareLock on relation 1259 of
database 35239378; blocked by process 28182.
Process 28182 waits for AccessShareLock on relation 1247 of database
35239378; blocked by process 28183.
HINT: �See server log for query details.

------------------------------
/var/log/postgresql-9.6
------------------------------
...
LOG: �received SIGHUP, reloading configuration files
ERROR: �canceling autovacuum task
CONTEXT: �automatic vacuum of table "mydatabase_test.someschema.sometable"
ERROR: �canceling autovacuum task
CONTEXT: �automatic vacuum of table "mydatabase_test.someschema.sometable"
ERROR: �canceling autovacuum task
CONTEXT: �automatic vacuum of table
"mydatabase_test.pg_toast.pg_toast_35240266"
LOG: �could not receive data from client: Connection timed out
LOG: �received SIGHUP, reloading configuration files
LOG: �received SIGHUP, reloading configuration files
ERROR: �canceling autovacuum task
CONTEXT: �automatic vacuum of table "mydatabase_test.someschema.sometable"
LOG: �received SIGHUP, reloading configuration files
LOG: �received SIGHUP, reloading configuration files
ERROR: �canceling autovacuum task
CONTEXT: �automatic vacuum of table
"mydatabase_test.pg_toast.pg_toast_35240330"
ERROR: �canceling autovacuum task
CONTEXT: �automatic vacuum of table "mydatabase_test.someschema.sometable"
ERROR: �deadlock detected
DETAIL: �Process 28183 waits for AccessShareLock on relation 1259 of
database 35239378; blocked by process 28182.
Process 28182 waits for AccessShareLock on relation 1247 of database
35239378; blocked by process 28183.
Process 28183: VACUUM (FULL) pg_catalog.pg_type;
Process 28182: VACUUM (FULL) pg_catalog.pg_class;
HINT: �See server log for query details.
STATEMENT: �VACUUM (FULL) pg_catalog.pg_type;
ERROR: �canceling statement due to user request
STATEMENT: �VACUUM (FULL) pg_catalog.pg_class;
ERROR: �canceling statement due to user request
STATEMENT: �VACUUM (FULL) pg_catalog.pg_depend;
ERROR: �canceling statement due to user request
STATEMENT: �VACUUM (FULL) someschema.sometable
LOG: �could not send data to client: Broken pipe
FATAL: �connection to client lost
LOG: �could not send data to client: Broken pipe
FATAL: �connection to client lost
ERROR: �canceling statement due to user request
STATEMENT: �VACUUM (FULL) mycompanydata.wildcard_replacement_bkp;
LOG: �could not send data to client: Broken pipe
STATEMENT: �VACUUM (FULL) mycompanydata.wildcard_replacement_bkp;
FATAL: �connection to client lost
ERROR: �canceling statement due to user request
STATEMENT: �VACUUM (FULL) pg_catalog.pg_proc;
LOG: �could not send data to client: Broken pipe
STATEMENT: �VACUUM (FULL) pg_catalog.pg_proc;
FATAL: �connection to client lost
ERROR: �canceling statement due to user request
STATEMENT: �VACUUM (FULL) mycompanydata.cpc_tag_score2;
LOG: �could not send data to client: Broken pipe
STATEMENT: �VACUUM (FULL) mycompanydata.cpc_tag_score2;
LOG: �could not send data to client: Broken pipe
FATAL: �connection to client lost
FATAL: �connection to client lost
LOG: �received SIGHUP, reloading configuration files
LOG: �received SIGHUP, reloading configuration files

--
Adrian Klaver
adrian.klaver@aklaver.com

@Adrian thanks.

I did a test vacuum full the "normal" non-parallel way (VACUUM FULL sql) previously on a test db copy and saw the DB size (postgres 9.6) shrink from 6.8 TB to 4.4 TB of actual, real disk space usage ($ df -d 1).

I don't know the reason so much space was "locked up" (other than there is a lot of data "churn" from daily updates). But I definitely do need to do the vac full on the production db to get down to the smaller size - cannot afford the 2.4 TB of "wasted" space on an ongoing basis.

Based on your email it sounds like the vacuuming of those pg_catalog.* tables is done regardless, as a normal part of doing vac full on my own database.

Unfortunately I still don't see an ideal way to run vacuum full in parallel via vacuumdb without running into the expected and documented deadlock. Only method I'm aware of is to list each table individually with "-t table1 -t table2..." to "vacuum db --jobs" which is not pleasant and not exceedingly beautiful.

Thanks.

On 5/14/20 11:27 AM, Eduard Rozenberg wrote:

@Adrian thanks.

I did a test vacuum full the "normal" non-parallel way (VACUUM FULL sql)
previously on a test db copy and saw the DB size (postgres 9.6) shrink
from 6.8 TB to 4.4 TB of actual, real disk space usage ($ df -d 1).

I don't know the reason so much space was "locked up" (other than there
is a lot of data "churn" from daily updates). But I definitely do need
to do the vac full on the production db to get down to the smaller size
- cannot afford the 2.4 TB of "wasted" space on an ongoing basis.

It may not be wasted space. A regular VACUUM marks space within a table
available for reuse(and not returned to OS) when it removes unneeded
tuples. It then fills that space up with new tuples, roughly speaking.
So if the vacuuming is working properly you will reach a steady state
where space is reused and the database on disk size grows slowly as
reusable space is occupied. I would monitor the database size on a
regular basis. My guess is that the VACUUM FULL is dropping the OS used
space and then it fills up again as the database does those updates.

Based on your email it sounds like the vacuuming of those pg_catalog.*
tables is done regardless, as a normal part of doing vac full on my own
database.

Unfortunately I still don't see an ideal way to run vacuum full in
parallel via vacuumdb without running into the expected and documented
deadlock. Only method I'm aware of is to list each table individually
with "-t table1 -t table2..." to "vacuum db --jobs" which is not
pleasant and not exceedingly beautiful.

Thanks.

--
Adrian Klaver
adrian.klaver@aklaver.com

@Adrian thanks again.

I read the postgres docs the same way - that previously used space is marked as available and therefore no need for vacuum full. Previously used = now available space, which gets re-used, in theory.

And yet this same DB is growing at 30-50 GB weekly, despite the fact that 2.4 TB of previously used space is clearly available ("clearly available" as proven by vacuum full shrinking the DB space usage by 2.4 TB). I did verify postgresql.conf has always been properly configured re: autovacuum: 'autovacuum = on'and 'track_counts = on'

I'm not planning on running VACUUM FULL regularly, just "this one time". And I was trying to to parallelize VACUUM FULL and minimize downtime of the production DB caused by table locking. And then I found the option of using "vacuumdb --jobs" which sounded like the perfect solution except for "well you can't actually use --jobs because you'll run into a deadlock and everybody knows that and nobody has a (good) solution for it" :).

--Ed

On 5/14/20 12:03 PM, Eduard Rozenberg wrote:

@Adrian thanks again.

I read the postgres docs the same way - that previously used space is marked as available and therefore no need for vacuum full. Previously used = now available space, which gets re-used, in theory.

And yet this same DB is growing at 30-50 GB weekly, despite the fact that 2.4 TB of previously used space is clearly available ("clearly available" as proven by vacuum full shrinking the DB space usage by 2.4 TB). I did verify postgresql.conf has always been properly configured re: autovacuum: 'autovacuum = on'and 'track_counts = on'

Well if I'm counting zeros right 50GB on 4.4TB database is ~1.14%. Does
that sound right for new data being added?

One place to look to see how aggressively the autovacuum is being done
here:

https://www.postgresql.org/docs/12/monitoring-stats.html#PG-STAT-ALL-TABLES-VIEW

What you find there may mean tweaking the values as explained here:

https://www.postgresql.org/docs/12/routine-vacuuming.html#AUTOVACUUM

Could be there are just a few tables that account for most of the churn
and a manual VACUUM on them is needed. Say if there are regularly
scheduled large UPDATEs to tables, incorporate a VACUUM after.

I'm not planning on running VACUUM FULL regularly, just "this one time". And I was trying to to parallelize VACUUM FULL and minimize downtime of the production DB caused by table locking. And then I found the option of using "vacuumdb --jobs" which sounded like the perfect solution except for "well you can't actually use --jobs because you'll run into a deadlock and everybody knows that and nobody has a (good) solution for it" :).

--Ed

On May 14, 2020, at 11:46, Adrian Klaver <adrian.klaver@aklaver.com> wrote:

On 5/14/20 11:27 AM, Eduard Rozenberg wrote:

@Adrian thanks.
I did a test vacuum full the "normal" non-parallel way (VACUUM FULL sql) previously on a test db copy and saw the DB size (postgres 9.6) shrink from 6.8 TB to 4.4 TB of actual, real disk space usage ($ df -d 1).
I don't know the reason so much space was "locked up" (other than there is a lot of data "churn" from daily updates). But I definitely do need to do the vac full on the production db to get down to the smaller size - cannot afford the 2.4 TB of "wasted" space on an ongoing basis.

It may not be wasted space. A regular VACUUM marks space within a table available for reuse(and not returned to OS) when it removes unneeded tuples. It then fills that space up with new tuples, roughly speaking. So if the vacuuming is working properly you will reach a steady state where space is reused and the database on disk size grows slowly as reusable space is occupied. I would monitor the database size on a regular basis. My guess is that the VACUUM FULL is dropping the OS used space and then it fills up again as the database does those updates.

Based on your email it sounds like the vacuuming of those pg_catalog.* tables is done regardless, as a normal part of doing vac full on my own database.
Unfortunately I still don't see an ideal way to run vacuum full in parallel via vacuumdb without running into the expected and documented deadlock. Only method I'm aware of is to list each table individually with "-t table1 -t table2..." to "vacuum db --jobs" which is not pleasant and not exceedingly beautiful.
Thanks.

--
Adrian Klaver
adrian.klaver@aklaver.com

--
Adrian Klaver
adrian.klaver@aklaver.com

You might have index fragmentation and possibly reindexing them conncurrently on PG12, should do that. As everyone agreed most of space will be marked for re-use later for table segments but indices in your case could be problem. On previous versions you can use pg_repack to do index rebuilds which will help reduce space consumed.

Regards,
Virendra

On Thursday, May 14, 2020, 1:20:41 PM PDT, Eduard Rozenberg <edrozenberg@gmail.com> wrote:

@Adrian thanks again.

I read the postgres docs the same way - that previously used space is marked as available and therefore no need for vacuum full. Previously used = now available space, which gets re-used, in theory.

And yet this same DB is growing at 30-50 GB weekly, despite the fact that 2.4 TB of previously used space is clearly available ("clearly available" as proven by vacuum full shrinking the DB space usage by 2.4 TB). I did verify postgresql.conf has always been properly configured re: autovacuum:  'autovacuum = on'and 'track_counts = on'

I'm not planning on running VACUUM FULL regularly, just "this one time". And I was trying to to parallelize VACUUM FULL and minimize downtime of the production DB caused by table locking. And then I found the option of using "vacuumdb --jobs" which sounded like the perfect solution except for "well you can't actually use --jobs because you'll run into a deadlock and everybody knows that and nobody has a (good) solution for it" :).

--Ed

On Thu, May 14, 2020 at 2:20 PM Eduard Rozenberg <edrozenberg@gmail.com>
wrote:

I did verify postgresql.conf has always been properly configured re:
autovacuum: 'autovacuum = on'and 'track_counts = on'

This may be insufficient to keep up if you have large tables. The default
scale factor allows for 20% of the rows to be dead before the autovacuum
will kick in to mark the space as available for reuse. Assuming you have
the I/O capacity and prefer to do a little cleanup more often rather than
HUGE cleanup work all at once on rare occasions, it may be ideal to look at
turning down the autovacuum_vacuum_scale_factor. You can tweak these
settings on large tables only, or increase the autovacuum_vacuum_threshold
at the same time to compensate a bit for decreasing the scale factor. You
can also look at pg_stat_activity for autovacuums, and if you see that some
are running for hours, then probably they are trying to do too much work
all at once and waiting too long before tidying up. Also, the default
autovacuum_vacuum_cost_delay was changed from 20ms to 2ms with PG12 so that
may be worth considering as a best practice even on older versions.