Hi,

Back in October 2009, I reported on strange warm-standby problems in
this thread: http://archives.postgresql.org/pgsql-admin/2009-10/msg00170. php

Just in case anyone still cares or is wondering, we found the problem.
The machine had bad RAM; we were getting undetected/uncorrected single-bit
errors creeping through! The thing that led to the discovery was a cron
job error complaining about a "SEHECT" statement when the Perl script
clearly read "SELECT". We swapped the RAM a while back and the problem
seems to have been cured.

Regards,

David.

--
Sent via pgsql-admin mailing list (pgsql-admin [at] postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin

On Wed, Jan 20, 2010 at 10:10:26AM -0500, David F. Skoll wrote:
> The machine had bad RAM; we were getting undetected/uncorrected single-bit
> errors creeping through!

who's the machine/memory vendor?

--
Sent via pgsql-admin mailing list (pgsql-admin [at] postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin

Ray Stell wrote:
>> The machine had bad RAM; we were getting undetected/uncorrected single-bit
>> errors creeping through!

> who's the machine/memory vendor?

I don't know exactly; it's a colocated machine that we don't own.
dmidecode (trimmed down) says:

System Information
Manufacturer: Supermicro
Product Name: C2SBM-Q
Version: 0123456789

[...]
Memory Device
Error Information Handle: No Error
Size: 2048 MB
Form Factor: DIMM
Bank Locator: DIMM 2
Type: DDR2
Type Detail: Synchronous
Speed: 800 MHz (1.2 ns)
Manufacturer: Kingston
Serial Number: 0DCC6845

But that's the good memory... I don't know what the bad memory was.

Regards,

David.

--
Sent via pgsql-admin mailing list (pgsql-admin [at] postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin