Re: Postgresql 8.2.4 on linux-sparc problem

Hello, Tom.

One again it happen.

postgres 11009     1       0 Jan12 ?        00:00:12 /usr/bin/
postmaster -D /data/idx/pgdata --silent-mode=true
postgres 11027 11009  0 Jan12 ?        00:26:55 postgres: logger process
postgres 11029 11009  0 Jan12 ?        00:00:21 postgres: writer process
postgres 11030 11009  0 Jan12 ?        00:16:50 postgres: stats
collector process
postgres 16751 11009  1 20:40 ?        00:00:12 postgres: stat stat
10.0.0.2(41239) idle
postgres 16753 11009  0 20:40 ?        00:00:11 postgres: stat stat
10.0.0.2(41244) idle
postgres 16758 11009  3 20:41 ?        00:00:35 postgres: stat stat
10.0.0.2(50546) SELECT
postgres 16760 11009  0 20:42 ?        00:00:00 postgres: stat stat
10.0.0.2(50573) idle
    postgres 16761 11009 99 20:42 ?        00:16:59 postgres: stat stat
10.0.0.2(50577) idle
postgres 16762 11009  0 20:43 ?        00:00:00 postgres: stat stat
10.0.0.2(50603) INSERT

I tried to use gdb but no success.

machupicchu ~ # gdb /usr/bin/postgres 16761
GNU gdb 6.6
Copyright (C) 2006 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and
you are
welcome to change it and/or distribute copies of it under certain
conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for
details.
This GDB was configured as "sparc-unknown-linux-gnu"...
Using host libthread_db library "/lib/libthread_db.so.1".
Attaching to program: /usr/bin/postgres, process 16761
....

I enabled  full logging to find a query after which pg die.  From
postgresql.conf:
log_directory = 'pg_log'
log_filename = 'postgresql-%Y-%m-%d_%H%M%S.log'
log_truncate_on_rotation = off
log_min_messages = error
log_min_error_statement = error
log_min_duration_statement = 1000
log_duration = on
log_line_prefix = '%m, %s, %r, %p, '
log_statement = 'all'

But I can't find nothing about PID 16761 or tcp port 50577 in logs! I
can find neighbour, but not this process.

Else:
# vmstat 1
procs -----------memory---------- ---swap-- -----io---- -system-- ----
cpu----
r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy
id wa
4 24   1408  25832 180800 3738664    0    0     5    23   31    5  8
1 84  6
4 24   1408  25832 180800 3738664    0    0     0     0 1010   14  0
50  0 50
4 24   1408  25832 180800 3738664    0    0     0     0 1005   11  0
50  0 50
4 24   1408  25832 180800 3738664    0    0     0     0 1005   11  0
50  0 50
4 24   1408  25832 180800 3738664    0    0     0     0 1004   11  0
50  0 50
4 24   1408  25832 180800 3738664    0    0     0     0 1005   13  0
50  0 50
4 24   1408  25832 180800 3738664    0    0     0     0 1004   13  0
50  0 50
4 24   1408  25832 180800 3738664    0    0     0     0 1004   11  0
50  0 50

Looks like problem is in IO-wait?

What to do? Where to dig?

In some weeks a want to migrate to 2 redundant sun fire v440 servers
with 4 storages. Main idea is to have HW redundancy. But now.. Don't
know what to say to my boss. HW is fine, but a lot of data loss.. :)

p.s. I commented my configuration for _mem options. Now it's default.