pg_upgrade 10.2

On 06/12/2018 01:34 PM, Murthy Nunna wrote:

In older versions of pg_upgrade (e.g from 9.2 to 9.3), I was able to run
pg_upgrade without stopping old cluster using the check flag.

pg_upgrade -b <old-bin> -B <new-bin> -d <old-data> -D <new-data> -p 5432
-P 5434 -r -v -k -c

Note the �c� flag at the end

I take the below to it mean it should work:

https://www.postgresql.org/docs/10/static/pgupgrade.html

"You can use pg_upgrade --check to perform only the checks, even if the

old server is still running. pg_upgrade --check will also outline any
manual adjustments you will need to make after the upgrade. If you are
going to be using link mode, you should use the --link option with
--check to enable link-mode-specific checks."

Might want to try without -k to see what happens.

More comments below.

However pg_upgrade in 10 (I tried from 9.3 to 10.4), when I did not stop
the old cluster, the upgrade failed:

***

There seems to be a postmaster servicing the old cluster.

Please shutdown that postmaster and try again.

Failure, exiting

Is this expected?

Also, when I stopped the old cluster and ran pg_upgrade with �-c� flag,
the file global/pg_control got renamed to global/pg_control.old. The
�-c� flag never renamed anything in the old cluster in older pg_upgrade

Again seems related to -k:

"
If you ran pg_upgrade without --link or did not start the new server,
the old cluster was not modified except that, if linking started, a .old
suffix was appended to $PGDATA/global/pg_control. To reuse the old
cluster, possibly remove the .old suffix from $PGDATA/global/pg_control;
you can then restart the old cluster.
"

--
Adrian Klaver
adrian.klaver@aklaver.com

Thanks Adrian.
I removed "-k" flag. But still got same error.

There seems to be a postmaster servicing the old cluster.
Please shutdown that postmaster and try again.
Failure, exiting

-----Original Message-----
From: Adrian Klaver [mailto:adrian.klaver@aklaver.com]
Sent: Tuesday, June 12, 2018 3:48 PM
To: Murthy Nunna <mnunna@fnal.gov>; pgsql-general@lists.postgresql.org; pgsql-admin@lists.postgresql.org; pgsql-performance@lists.postgresql.org
Subject: Re: pg_upgrade 10.2

On 06/12/2018 01:34 PM, Murthy Nunna wrote:

In older versions of pg_upgrade (e.g from 9.2 to 9.3), I was able to
run pg_upgrade without stopping old cluster using the check flag.

pg_upgrade -b <old-bin> -B <new-bin> -d <old-data> -D <new-data> -p
5432 -P 5434 -r -v -k -c

Note the "c" flag at the end

I take the below to it mean it should work:

https://urldefense.proofpoint.com/v2/url?u=https-3A__www.postgresql.org_docs_10_static_pgupgrade.html&amp;d=DwID-g&amp;c=gRgGjJ3BkIsb5y6s49QqsA&amp;r=0wrsmPzpZSao0v32yCcG2Q&amp;m=g2e1NMngBLIcEgi5UjlCHkyJ5zK1Su-vsaRw0Y9N0Dc&amp;s=PDVmjA_uW6cJvV4lWR8vgkiArplzgd5Rs4taLA6ZY6Q&amp;e=

"You can use pg_upgrade --check to perform only the checks, even if
the

old server is still running. pg_upgrade --check will also outline any manual adjustments you will need to make after the upgrade. If you are going to be using link mode, you should use the --link option with --check to enable link-mode-specific checks."

Might want to try without -k to see what happens.

More comments below.

However pg_upgrade in 10 (I tried from 9.3 to 10.4), when I did not
stop the old cluster, the upgrade failed:

***

There seems to be a postmaster servicing the old cluster.

Please shutdown that postmaster and try again.

Failure, exiting

Is this expected?

Also, when I stopped the old cluster and ran pg_upgrade with "-c"
flag, the file global/pg_control got renamed to global/pg_control.old.
The "-c" flag never renamed anything in the old cluster in older
pg_upgrade

Again seems related to -k:

"
If you ran pg_upgrade without --link or did not start the new server, the old cluster was not modified except that, if linking started, a .old suffix was appended to $PGDATA/global/pg_control. To reuse the old cluster, possibly remove the .old suffix from $PGDATA/global/pg_control; you can then restart the old cluster.
"

--
Adrian Klaver
adrian.klaver@aklaver.com

On 06/12/2018 01:58 PM, Murthy Nunna wrote:

Thanks Adrian.
I removed "-k" flag. But still got same error.

There seems to be a postmaster servicing the old cluster.
Please shutdown that postmaster and try again.
Failure, exiting

Well according to the code in pg_upgrade.c that message should not be
reached when the check option is specified:

if (!user_opts.check)
pg_fatal("There seems to be a postmaster servicing the old
cluster.\n"
"Please shutdown that postmaster and try again.\n");
else
*live_check = true;

Can we see the actual command you ran?

--
Adrian Klaver
adrian.klaver@aklaver.com

pg_upgrade -V
pg_upgrade (PostgreSQL) 10.4

pg_upgrade -b /fnal/ups/prd/postgres/v9_3_14_x64/Linux-2-6/bin -B /fnal/ups/prd/postgres/v10_4_x64/Linux-2-6/bin -d /data0/pgdata/ifb_prd_last -D /data0/pgdata/ifb_prd_last_104 -p 5433 -P 5434 -r -v –c

-----Original Message-----
From: Adrian Klaver [mailto:adrian.klaver@aklaver.com]
Sent: Tuesday, June 12, 2018 4:13 PM
To: Murthy Nunna <mnunna@fnal.gov>; pgsql-general@lists.postgresql.org; pgsql-admin@lists.postgresql.org; pgsql-performance@lists.postgresql.org
Subject: Re: pg_upgrade 10.2

On 06/12/2018 01:58 PM, Murthy Nunna wrote:

Thanks Adrian.
I removed "-k" flag. But still got same error.

There seems to be a postmaster servicing the old cluster.
Please shutdown that postmaster and try again.
Failure, exiting

Well according to the code in pg_upgrade.c that message should not be reached when the check option is specified:

if (!user_opts.check)
pg_fatal("There seems to be a postmaster servicing the old cluster.\n"
"Please shutdown that postmaster and try again.\n"); else
*live_check = true;

Can we see the actual command you ran?

--
Adrian Klaver
adrian.klaver@aklaver.com

On 06/12/2018 02:18 PM, Murthy Nunna wrote:

pg_upgrade -V
pg_upgrade (PostgreSQL) 10.4

pg_upgrade -b /fnal/ups/prd/postgres/v9_3_14_x64/Linux-2-6/bin -B /fnal/ups/prd/postgres/v10_4_x64/Linux-2-6/bin -d /data0/pgdata/ifb_prd_last -D /data0/pgdata/ifb_prd_last_104 -p 5433 -P 5434 -r -v –c

Looks good to me. The only thing that stands out is that in your
original post you had:

-p 5432

and above you have:

-p 5433

Not sure if that makes a difference.

The only suggestion I have at the moment is to move -c from the end of
the line to somewhere earlier on the chance that there is a bug that is
not finding it when it's at the end.

--
Adrian Klaver
adrian.klaver@aklaver.com

Hi Adrian,

Port numbers are correct.

I moved the position of -c (-p 5433 -P 5434 -c -r -v). Now it is NOT complaining about old cluster running. However, I am running into a different problem.

New cluster database "ifb_prd_last" is not empty
Failure, exiting

Note: ifb_prd_last is not new cluster. It is actually old cluster.

Is this possibly because in one of my earlier attempts where I shutdown old cluster and ran pg_upgrade with -c at the end of the command line. I think -c was ignored and my cluster has been upgraded in that attempt. Is that possible?

-----Original Message-----
From: Adrian Klaver [mailto:adrian.klaver@aklaver.com]
Sent: Tuesday, June 12, 2018 4:35 PM
To: Murthy Nunna <mnunna@fnal.gov>; pgsql-general@lists.postgresql.org; pgsql-admin@lists.postgresql.org; pgsql-performance@lists.postgresql.org
Subject: Re: pg_upgrade 10.2

On 06/12/2018 02:18 PM, Murthy Nunna wrote:

pg_upgrade -V
pg_upgrade (PostgreSQL) 10.4

pg_upgrade -b /fnal/ups/prd/postgres/v9_3_14_x64/Linux-2-6/bin -B
/fnal/ups/prd/postgres/v10_4_x64/Linux-2-6/bin -d
/data0/pgdata/ifb_prd_last -D /data0/pgdata/ifb_prd_last_104 -p 5433
-P 5434 -r -v –c

Looks good to me. The only thing that stands out is that in your original post you had:

-p 5432

and above you have:

-p 5433

Not sure if that makes a difference.

The only suggestion I have at the moment is to move -c from the end of the line to somewhere earlier on the chance that there is a bug that is not finding it when it's at the end.

--
Adrian Klaver
adrian.klaver@aklaver.com

Murthy Nunna <mnunna@fnal.gov> writes:

Hi Adrian,

Port numbers are correct.

I moved the position of -c (-p 5433 -P 5434 -c -r -v). Now it is NOT complaining about old cluster running. However, I am running into a different problem.

I noted in your earlier message the final -c... the dash was not a
regular 7bit ascii char but some UTF or whatever dash char.

I wonder if that's what you fed your shell and it caused a silent
parsing issue, eg the -c dropped.

But of course email clients wrap and mangle text like that all sorts of
fun ways so lordy knows just what you originally sent :-)

FWIW

New cluster database "ifb_prd_last" is not empty
Failure, exiting

Note: ifb_prd_last is not new cluster. It is actually old cluster.

Is this possibly because in one of my earlier attempts where I
shutdown old cluster and ran pg_upgrade with -c at the end of the
command line. I think -c was ignored and my cluster has been upgraded
in that attempt. Is that possible?

-----Original Message-----
From: Adrian Klaver [mailto:adrian.klaver@aklaver.com]
Sent: Tuesday, June 12, 2018 4:35 PM
To: Murthy Nunna <mnunna@fnal.gov>; pgsql-general@lists.postgresql.org; pgsql-admin@lists.postgresql.org; pgsql-performance@lists.postgresql.org
Subject: Re: pg_upgrade 10.2

On 06/12/2018 02:18 PM, Murthy Nunna wrote:

pg_upgrade -V
pg_upgrade (PostgreSQL) 10.4

pg_upgrade -b /fnal/ups/prd/postgres/v9_3_14_x64/Linux-2-6/bin -B
/fnal/ups/prd/postgres/v10_4_x64/Linux-2-6/bin -d
/data0/pgdata/ifb_prd_last -D /data0/pgdata/ifb_prd_last_104 -p 5433
-P 5434 -r -v –c

Looks good to me. The only thing that stands out is that in your original post you had:

-p 5432

and above you have:

-p 5433

Not sure if that makes a difference.

The only suggestion I have at the moment is to move -c from the end of the line to somewhere earlier on the chance that there is a bug that is not finding it when it's at the end.

--
Adrian Klaver
adrian.klaver@aklaver.com

--
Jerry Sievers
Postgres DBA/Development Consulting
e: postgres.consulting@comcast.net
p: 312.241.7800

Jerry,

OMG, I think you nailed this... I know what I did. I cut/pasted the command from an e-mail... I have seen this issue before with stuff not related to postgres. But then those commands failed in syntax error and then you know what you did wrong.

Similarly, I expect pg_upgrade to throw an error if it finds something it doesn't understand instead of ignoring and causing damage. Don't you agree?

Thanks for pointing that out. I will redo my upgrade.

-r -v -k -c --- good flags no utf8
-r -v -k –c --- bad flags....

-----Original Message-----
From: Jerry Sievers [mailto:gsievers19@comcast.net]
Sent: Tuesday, June 12, 2018 6:24 PM
To: Murthy Nunna <mnunna@fnal.gov>
Cc: Adrian Klaver <adrian.klaver@aklaver.com>; pgsql-general@lists.postgresql.org; pgsql-admin@lists.postgresql.org; pgsql-performance@lists.postgresql.org
Subject: Re: pg_upgrade 10.2

Murthy Nunna <mnunna@fnal.gov> writes:

Hi Adrian,

Port numbers are correct.

I moved the position of -c (-p 5433 -P 5434 -c -r -v). Now it is NOT complaining about old cluster running. However, I am running into a different problem.

I noted in your earlier message the final -c... the dash was not a regular 7bit ascii char but some UTF or whatever dash char.

I wonder if that's what you fed your shell and it caused a silent parsing issue, eg the -c dropped.

But of course email clients wrap and mangle text like that all sorts of fun ways so lordy knows just what you originally sent :-)

FWIW

New cluster database "ifb_prd_last" is not empty Failure, exiting

Note: ifb_prd_last is not new cluster. It is actually old cluster.

Is this possibly because in one of my earlier attempts where I
shutdown old cluster and ran pg_upgrade with -c at the end of the
command line. I think -c was ignored and my cluster has been upgraded
in that attempt. Is that possible?

-----Original Message-----
From: Adrian Klaver [mailto:adrian.klaver@aklaver.com]
Sent: Tuesday, June 12, 2018 4:35 PM
To: Murthy Nunna <mnunna@fnal.gov>;
pgsql-general@lists.postgresql.org; pgsql-admin@lists.postgresql.org;
pgsql-performance@lists.postgresql.org
Subject: Re: pg_upgrade 10.2

On 06/12/2018 02:18 PM, Murthy Nunna wrote:

pg_upgrade -V
pg_upgrade (PostgreSQL) 10.4

pg_upgrade -b /fnal/ups/prd/postgres/v9_3_14_x64/Linux-2-6/bin -B
/fnal/ups/prd/postgres/v10_4_x64/Linux-2-6/bin -d
/data0/pgdata/ifb_prd_last -D /data0/pgdata/ifb_prd_last_104 -p 5433
-P 5434 -r -v –c

Looks good to me. The only thing that stands out is that in your original post you had:

-p 5432

and above you have:

-p 5433

Not sure if that makes a difference.

The only suggestion I have at the moment is to move -c from the end of the line to somewhere earlier on the chance that there is a bug that is not finding it when it's at the end.

--
Adrian Klaver
adrian.klaver@aklaver.com

--
Jerry Sievers
Postgres DBA/Development Consulting
e: postgres.consulting@comcast.net
p: 312.241.7800

On 06/12/2018 02:49 PM, Murthy Nunna wrote:

Hi Adrian,

Port numbers are correct.

I moved the position of -c (-p 5433 -P 5434 -c -r -v). Now it is NOT complaining about old cluster running. However, I am running into a different problem.

New cluster database "ifb_prd_last" is not empty
Failure, exiting

Note: ifb_prd_last is not new cluster. It is actually old cluster.

Is this possibly because in one of my earlier attempts where I shutdown old cluster and ran pg_upgrade with -c at the end of the command line. I think -c was ignored and my cluster has been upgraded in that attempt. Is that possible?

I don't so because it exited before it got the upgrading part.

--
Adrian Klaver
adrian.klaver@aklaver.com

Murthy Nunna <mnunna@fnal.gov> writes:

Jerry,

OMG, I think you nailed this... I know what I did. I cut/pasted the
command from an e-mail... I have seen this issue before with stuff not

Oh! I suggest you lose that habit ASAP before ever issuing another
command to anything :-)

related to postgres. But then those commands failed in syntax error
and then you know what you did wrong.

Similarly, I expect pg_upgrade to throw an error if it finds something it doesn't understand instead of ignoring and causing damage. Don't you agree?

Well, pg_upgrade might never have seen your $silly-dash since possibly
your shell or terminal driver swallowed it.

Thanks for pointing that out. I will redo my upgrade.

-r -v -k -c --- good flags no utf8
-r -v -k –c --- bad flags....

-----Original Message-----
From: Jerry Sievers [mailto:gsievers19@comcast.net]
Sent: Tuesday, June 12, 2018 6:24 PM
To: Murthy Nunna <mnunna@fnal.gov>
Cc: Adrian Klaver <adrian.klaver@aklaver.com>; pgsql-general@lists.postgresql.org; pgsql-admin@lists.postgresql.org; pgsql-performance@lists.postgresql.org
Subject: Re: pg_upgrade 10.2

Murthy Nunna <mnunna@fnal.gov> writes:

Hi Adrian,

Port numbers are correct.

I moved the position of -c (-p 5433 -P 5434 -c -r -v). Now it is NOT complaining about old cluster running. However, I am running into a different problem.

I noted in your earlier message the final -c... the dash was not a regular 7bit ascii char but some UTF or whatever dash char.

I wonder if that's what you fed your shell and it caused a silent parsing issue, eg the -c dropped.

But of course email clients wrap and mangle text like that all sorts of fun ways so lordy knows just what you originally sent :-)

FWIW

New cluster database "ifb_prd_last" is not empty Failure, exiting

Note: ifb_prd_last is not new cluster. It is actually old cluster.

Is this possibly because in one of my earlier attempts where I
shutdown old cluster and ran pg_upgrade with -c at the end of the
command line. I think -c was ignored and my cluster has been upgraded
in that attempt. Is that possible?

-----Original Message-----
From: Adrian Klaver [mailto:adrian.klaver@aklaver.com]
Sent: Tuesday, June 12, 2018 4:35 PM
To: Murthy Nunna <mnunna@fnal.gov>;
pgsql-general@lists.postgresql.org; pgsql-admin@lists.postgresql.org;
pgsql-performance@lists.postgresql.org
Subject: Re: pg_upgrade 10.2

On 06/12/2018 02:18 PM, Murthy Nunna wrote:

pg_upgrade -V
pg_upgrade (PostgreSQL) 10.4

pg_upgrade -b /fnal/ups/prd/postgres/v9_3_14_x64/Linux-2-6/bin -B
/fnal/ups/prd/postgres/v10_4_x64/Linux-2-6/bin -d
/data0/pgdata/ifb_prd_last -D /data0/pgdata/ifb_prd_last_104 -p 5433
-P 5434 -r -v –c

Looks good to me. The only thing that stands out is that in your original post you had:

-p 5432

and above you have:

-p 5433

Not sure if that makes a difference.

The only suggestion I have at the moment is to move -c from the end of the line to somewhere earlier on the chance that there is a bug that is not finding it when it's at the end.

--
Adrian Klaver
adrian.klaver@aklaver.com

--
Jerry Sievers
Postgres DBA/Development Consulting
e: postgres.consulting@comcast.net
p: 312.241.7800

--
Jerry Sievers
Postgres DBA/Development Consulting
e: postgres.consulting@comcast.net
p: 312.241.7800

Murthy Nunna <mnunna@fnal.gov> writes:

<snip>

BTW, this message was and remained cross-posted to 3 groups which is
considered bad style around here and I was negligent too in the previous
reply which also went out to all of them.

Please take note.

Thank

--
Jerry Sievers
Postgres DBA/Development Consulting
e: postgres.consulting@comcast.net
p: 312.241.7800