Slony1 PostgreSQL Replication

Author

SYNC errors

Ian Burrell

2005-07-20, 8:25 pm

We are getting these errors every time a SYNC tries to happen:

2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: "start
transaction; set enable_seqscan =3D off; set enable_indexscan
=3D on; " PGRES_FATAL_ERROR=20

2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: "close LOG; "
PGRES_FATAL_ERROR
2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: "rollback
transaction; set enable_seqscan =3D default; set enable_indexscan =3D
default; " PGRES_FATAL_ERROR=20

2005-07-20 12:00:42 PDT DEBUG2 remoteHelperThread_1
_1: 209032.740
seconds until close cursor
2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: helper 1 finished
with error
2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: SYNC aborted

Unfortunately, it is not printing out the error from PostgreSQL which
makes it hard to figure out why it is failing. Also, I added line
breaks after all the PGRES_FATAL_ERROR because they were all wrapped
together.

Does anyone have an idea what is calling the failure?

- Ian

Ian Burrell

2005-07-20, 8:25 pm

I think I found the cause of the problem. The errors started with:

2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: "start
transaction; set enable_seqscan =3D off; set enable_indexscan
=3D on; " PGRES_FATAL_ERROR server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: "close LOG; "
PGRES_FATAL_ERROR
2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: "rollback
transaction; set enable_seqscan =3D default; set enable_indexscan =3D
default; " PGRES_FATAL_ERROR
2005-07-19 15:17:23 PDT DEBUG2 remoteHelperThread_1
_1: 134433.825
seconds until close cursor
2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: helper 1 finished
with error
2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: SYNC aborted

I am guessing that the backend or the connection died. The SYNC's
kept failing with the message below. It is bad that slon did not
detect that the connection had failed and did not abort. This was
only fixed by restarting slon. What is weird is there was nothing in
the postgres server logs about this.

- Ian

On 7/20/05, Ian Burrell <ianburrell- Re5JQEeQqe8AvxtiuMwx
3w@public.gmane.org> wrote:
> We are getting these errors every time a SYNC tries to happen:
>=20
> 2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: "start
> transaction; set enable_seqscan =3D off; set enable_indexscan
> =3D on; " PGRES_FATAL_ERROR
> 2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: "close LOG; "
> PGRES_FATAL_ERROR
> 2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: "rollback
> transaction; set enable_seqscan =3D default; set enable_indexscan =3D
> default; " PGRES_FATAL_ERROR
> 2005-07-20 12:00:42 PDT DEBUG2 remoteHelperThread_1
_1: 209032.740
> seconds until close cursor
> 2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: helper 1 finished
> with error
> 2005-07-20 12:00:42 PDT ERROR remoteWorkerThread_1
: SYNC aborted
>=20
> Unfortunately, it is not printing out the error from PostgreSQL which
> makes it hard to figure out why it is failing. Also, I added line
> breaks after all the PGRES_FATAL_ERROR because they were all wrapped
> together.
>=20
> Does anyone have an idea what is calling the failure?
>=20
> - Ian
>

Jan Wieck

2005-07-20, 8:25 pm

Something causes the postgres backend to coredump. You should find a
core in the database directory. Please get a backtrace of that with a
debugger.

Jan

On 7/20/2005 3:19 PM, Ian Burrell wrote:
> I think I found the cause of the problem. The errors started with:
>
> 2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: "start
> transaction; set enable_seqscan = off; set enable_indexscan
> = on; " PGRES_FATAL_ERROR server closed the connection unexpectedly
> This probably means the server terminated abnormally
> before or while processing the request.
> 2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: "close LOG; "
> PGRES_FATAL_ERROR
> 2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: "rollback
> transaction; set enable_seqscan = default; set enable_indexscan =
> default; " PGRES_FATAL_ERROR
> 2005-07-19 15:17:23 PDT DEBUG2 remoteHelperThread_1
_1: 134433.825
> seconds until close cursor
> 2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: helper 1 finished
> with error
> 2005-07-19 15:17:23 PDT ERROR remoteWorkerThread_1
: SYNC aborted
>
> I am guessing that the backend or the connection died. The SYNC's
> kept failing with the message below. It is bad that slon did not
> detect that the connection had failed and did not abort. This was
> only fixed by restarting slon. What is weird is there was nothing in
> the postgres server logs about this.
>
> - Ian
>
> On 7/20/05, Ian Burrell <ianburrell- Re5JQEeQqe8AvxtiuMwx
3w@public.gmane.org> wrote:
> ____________________
____________________
_______
> Slony1-general mailing list
> Slony1-general- AuKwsB3Fm+ugFIWk8tvy
RWD2FQJk+8+b@public.gmane.org
> http://gborg.postgresql.org/mailman.../slony1-general

--
#===================
====================
====================
===========#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#===================
====================
=========== JanWieck- bwPqjjyvM7QAvxtiuMwx
3w@public.gmane.org #

Sponsored Links