anjali_524 <anjali_524@yahoo.co.in> writes:
> I am using Cent OS 6 () and Postgresql 9.0.4. I was running Enterprise setup
> of 4 node cluster. I got following error in postgres
> Reading symbols from /opt/openkaz/lib/kazsql.so...done.
> Loaded symbols for /opt/openkaz/lib/kazsql.so
> Core was generated by `postgres: kazeon KazDB 192.168.50.131(37625) SELECT
> '.
> Program terminated with signal 11, Segmentation fault.
> #0 0x00007f63bc79e0b8 in memcpy () from /lib64/libc.so.6
> Missing separate debuginfos, use: debuginfo-install
> kazsqldb-1.0.2-9.0.4.x86_64
> (gdb) bt
> #0 0x00007f63bc79e0b8 in memcpy () from /lib64/libc.so.6
> #1 0x000000000044d989 in heap_fill_tuple ()
> #2 0x000000000044dd91 in heap_form_minimal_tuple ()
> #3 0x00000000006ce356 in copytup_heap ()
> #4 0x00000000006ccfd1 in tuplesort_puttupleslot ()
> #5 0x000000000055e591 in ExecSort ()
> #6 0x00000000005480b8 in ExecProcNode ()
> #7 0x0000000000554e62 in ExecAgg ()
> #8 0x0000000000548098 in ExecProcNode ()
> #9 0x0000000000546f32 in standard_ExecutorRun ()
> #10 0x00000000005fc517 in PortalRunSelect ()
> #11 0x00000000005fd788 in PortalRun ()
> #12 0x00000000005fb223 in PostgresMain ()
> #13 0x00000000005c89e9 in ServerLoop ()
> #14 0x00000000005cafbc in PostmasterMain ()
> #15 0x0000000000572d98 in main ()
Apparently the Sort step is trying to copy an input tuple that is
corrupted somehow. That could be because of corrupted data on-disk,
or there could be a software bug that mangled the tuple during an
earlier processing step within this query.
Can you pg_dump the table(s) this query was reading without error?
If not, then it's on-disk data corruption. If so, then it's a
processing bug, but we'd need sufficient information to reproduce
the problem before we can fix it.
BTW, 9.0.4 is over a year old --- you might consider updating to
the current 9.0.x release, which is 9.0.9. Also, it sounds like
you might be using EDB's commercial version of Postgres, in which
case you really ought to go to them for support.
regards, tom lane
