Housekeeping - updating cache fails

[sam_vimes]

PT3 runs in a virtualbox VM. PostgreSQL 8.4 native in Linux.
Everything worked fine.
Yesterday I imported 250k hands and now everytime I run "update cache" it fails. The time it takes to show me the error messages varies. Sometimes it fails after 5 minutes, sometimes after over an hour.
The error message is:"An error occurred in the internal thread.
Exception: Connection Lost to Database: PT3_2010_02_05_110433".
The postgreSQL log looks like this after an attempted cache update:
"2010-03-02 11:19:59 CET WARNING: there is already a transaction in progress
2010-03-02 11:19:59 CET WARNING: there is already a transaction in progress
2010-03-02 11:19:59 CET WARNING: there is already a transaction in progress
2010-03-02 11:20:00 CET WARNING: there is already a transaction in progress
2010-03-02 11:20:01 CET WARNING: there is already a transaction in progress
2010-03-02 11:20:01 CET WARNING: there is already a transaction in progress
2010-03-02 11:20:01 CET LOG: unexpected EOF on client connection"

I have no firewall installed and the windows firewall is disabled.

[sam_vimes]

and the pokertracker.log looks like this:

2010/03/02 06:19:27.556: [00000D70][StatManager] Cache update is 18 percent complete (2,651/14,725).
2010/03/02 06:20:32.399: [00000D70][StatManager] Cache update is 19 percent complete (2,798/14,725).
2010/03/02 06:21:57.682: [00000D70][PostgresConnection (Error)] Unable to locate PostgreSQL information in the registry.
2010/03/02 06:21:57.682: [00000D70][PostgreConnection] Attempt to restart PostgreSQL failed: Unable to locate PostgreSQL information in the registry.
2010/03/02 06:21:57.682: [00000D70][PostgresConnection (Error)] Connection Lost to Database: PT3_2010_02_05_110433
2010/03/02 06:21:57.682: [00000D70][WindowDialogDatabaseHouseKeepProgress (Error)] Exception: Connection Lost to Database: PT3_2010_02_05_110433
2010/03/02 06:23:35.943: [- Main -][PostgresConnection] Connection lost, attempting to reset
2010/03/02 06:23:37.235: [- Main -][App] Shutting down subsystems...
2010/03/02 06:23:37.265: [- Main -][HudIpcInterface] Closing connection...
2010/03/02 06:23:37.345: [- Main -][System] Application exiting
2010/03/02 06:23:37.575: [- Main -][PostgresConnection] [0x0440b7e0] Disconnecting from database: PT3_2010_02_05_110433
2010/03/02 06:23:37.656: [- Main -][PostgresConnection] [0x0440b710] Disconnecting from database: PT3_2010_02_05_110433
2010/03/02 06:23:37.666: [- Main -][PostgresConnection] [0x0440b850] Disconnecting from database: PT3_2010_02_05_110433
2010/03/02 06:23:37.666: [- Main -][PostgresConnection] [0x04401700] Disconnecting from database: PT3_2010_02_05_110433

[kraada]

Try upgrading PostgreSQL to 8.3.9 and then try again - for large cache updates 8.3.4 has been known to crash and the problem is fixed in the later 8.3.x revisions.

[sam_vimes]

I'm using postgresql 8.4.2. So this would be a downgrade. No support for 8.4.x?

[kraada]

Officially not yet but this is the first I've heard of a PostgreSQL 8.4 crash of any sort relating to PT3.

Please report this problem with the log files so that we can investigate further for you.

[ninanu]

---WRONG THREAD--- should have been "Re: Beta 30.7 Update Cache issue"

My setup, PT in VM, Postgres 8.4.2 on Ubuntu 9.04. I've had a problem with updating cache sind installing Beta 30.7 that fits your description.

I didn't find any help in the forums about this, but there was something about Postgres Performance, so what I'm trying to say:

Update Cache worked now after I changed postgres.conf

shared_buffers = 256MB
work_mem = 32MB
logging_collector = off

This wouldn't work initially because the kernel.shmmax value was too low. I set it to (256+32)*1024*1024=301989888 (288MB in Bytes) via

"sysctl -w kernel.shmmax=301989888"

and I put it in sysctl.conf.

So with more memory allocated to the server, Update Cache worked.

Josiah could you please comment if this works for you at all since I'd like to know if this is indeed a fix (seems so right now for my setup).

[ninanu]

WRONG THREAD AGAIN

but update cache doesnt work again so no fix.

so both posts could be deleted.

[kraada]

Have you reported your problem per the link I gave you?