Re: Memory/Performance issue: server closed the connection

Tom Lane wrote:

>That shouldn't happen.  What shows up in the postmaster log?
>
Here is the output of the log:

2005-10-11 17:22:00 CEST (@) [] DEBUG:  checkpoint complete; 0
transaction log file(s) added, 0 removed, 3 recycled
2005-10-11 17:22:26 CEST (@) [] LOG:  checkpoints are occurring too
frequently (27 seconds apart)
2005-10-11 17:22:26 CEST (@) [] HINT:  Consider increasing the
configuration parameter "checkpoint_segments".
2005-10-11 17:22:26 CEST (@) [] DEBUG:  checkpoint starting
2005-10-11 17:22:27 CEST (@) [] DEBUG:  checkpoint complete; 0
transaction log file(s) added, 0 removed, 3 recycled
2005-10-11 17:22:28 CEST ([unknown]@[unknown]) [2005-10-11 17:22:28
CEST] LOG:  incomplete startup packet
2005-10-11 17:22:28 CEST (postgres@TheCore) [2005-10-11 17:08:49 CEST]
PANIC:  right sibling's left-link doesn't match
2005-10-11 17:22:28 CEST (postgres@TheCore) [2005-10-11 17:08:49 CEST]
STATEMENT:  update aaa set blinding_flag = 'D' where study_id = 144
2005-10-11 17:22:28 CEST (@) [] LOG:  server process (PID 13726) was
terminated by signal 6
2005-10-11 17:22:28 CEST (@) [] LOG:  terminating any other active
server processes
2005-10-11 17:22:28 CEST (postgres@TheCore) [2005-10-11 17:08:48 CEST]
WARNING:  terminating connection because of crash of another server process
2005-10-11 17:22:28 CEST (postgres@TheCore) [2005-10-11 17:08:48 CEST]
DETAIL:  The postmaster has commanded this server process to roll back
the current tran
saction and exit, because another server process exited abnormally and
possibly corrupted shared memory.
2005-10-11 17:22:28 CEST (postgres@TheCore) [2005-10-11 17:08:48 CEST]
HINT:  In a moment you should be able to reconnect to the database and
repeat your comm
and.
2005-10-11 17:22:28 CEST (postgres@template1) [2005-10-11 17:08:48 CEST]
WARNING:  terminating connection because of crash of another server process
2005-10-11 17:22:28 CEST (postgres@template1) [2005-10-11 17:08:48 CEST]
DETAIL:  The postmaster has commanded this server process to roll back
the current tr
ansaction and exit, because another server process exited abnormally and
possibly corrupted shared memory.
2005-10-11 17:22:28 CEST (postgres@template1) [2005-10-11 17:08:48 CEST]
HINT:  In a moment you should be able to reconnect to the database and
repeat your co
mmand.
2005-10-11 17:22:28 CEST (@) [] LOG:  all server processes terminated;
reinitializing
2005-10-11 17:22:28 CEST (@) [] LOG:  database system was interrupted at
2005-10-11 17:22:27 CEST
2005-10-11 17:22:28 CEST (@) [] LOG:  checkpoint record is at B/FB02FD7C
2005-10-11 17:22:28 CEST (@) [] LOG:  redo record is at B/FB008754; undo
record is at 0/0; shutdown FALSE
2005-10-11 17:22:28 CEST (@) [] LOG:  next transaction ID: 4546927; next
OID: 83401
2005-10-11 17:22:28 CEST (@) [] LOG:  database system was not properly
shut down; automatic recovery in progress
2005-10-11 17:22:28 CEST (@) [] LOG:  redo starts at B/FB008754
2005-10-11 17:22:28 CEST (@) [] LOG:  unexpected pageaddr B/F4150000 in
log file 11, segment 251, offset 1376256
2005-10-11 17:22:28 CEST (@) [] LOG:  redo done at B/FB14FF98
2005-10-11 17:22:28 CEST (@) [] LOG:  database system is ready
2005-10-11 17:22:48 CEST ([unknown]@[unknown]) [2005-10-11 17:22:48
CEST] LOG:  incomplete startup packet

> If the server is dumping core, can you get a stack trace from it?
>
>
I can't find any special file in /pgdb/system/base or its subdirectories.