I received this from a friend:
Lance Rushing wrote:
> Something bad happened in Postgres. I lost the ability to connect to the
> 'pegasus2' database.
>
> i was doing some updates when I got an error, saying something about the
> backend dying...
>
> When I restarted the database I could no longer connect to my database. I
> can connect to the other databases, just not 'pegasus2'.
This is a 7.3.3 installation on Red Hat (9 I think).
I don't have any more detail yet on exactly what he was doing at this
point, but I grabbed a copy of $PGDATA and looked at it on my own
machine (since he doesn't have debug and assert support). Logging into
any other database works fine, but the offending database produces this
backtrace:
...
#85355 0x0817c0b5 in RelationBuildDesc (buildinfo= {infotype = 2, i = {info_id = 135924865, info_name =
0x81a0c81
"pg_trigger"}}, oldrelation=0x0) at relcache.c:890
#85356 0x0817cf54 in RelationSysNameGetRelation (relationName=0x81a0c81
"pg_trigger") at relcache.c:1591
#85357 0x0807c066 in relation_openr (sysRelationName=0x81a0c81
"pg_trigger", lockmode=1) at heapam.c:550
#85358 0x0807c20a in heap_openr (sysRelationName=0x81a0c81 "pg_trigger",
lockmode=1) at heapam.c:660
#85359 0x080d5872 in RelationBuildTriggers (relation=0x41ad740c) at
trigger.c:720
#85360 0x0817c0b5 in RelationBuildDesc (buildinfo= {infotype = 2, i = {info_id = 135924865, info_name =
0x81a0c81
"pg_trigger"}}, oldrelation=0x0) at relcache.c:890
#85361 0x0817cf54 in RelationSysNameGetRelation (relationName=0x81a0c81
"pg_trigger") at relcache.c:1591
#85362 0x0807c066 in relation_openr (sysRelationName=0x81a0c81
"pg_trigger", lockmode=1) at heapam.c:550
#85363 0x0807c20a in heap_openr (sysRelationName=0x81a0c81 "pg_trigger",
lockmode=1) at heapam.c:660
#85364 0x080d5872 in RelationBuildTriggers (relation=0x41ad6b54) at
trigger.c:720
#85365 0x0817c0b5 in RelationBuildDesc (buildinfo= {infotype = 2, i = {info_id = 135924865, info_name =
0x81a0c81
"pg_trigger"}}, oldrelation=0x0) at relcache.c:890
...
i.e. it recurses itself to a horrible death.
Any ideas? Is this an indication that pg_trigger is corrupted?
Thanks,
Joe