Thread: Numerous postmaster processes after upgrading to 7.2.3

Numerous postmaster processes after upgrading to 7.2.3

From
Ericson Smith
Date:
Hi all,

Today we upgraded to ver 7.2.3 from 7.2.2.
I also dumped and restored the data before doing so, just in case :-)

I've noticed a few things...

1. The database server load is much higher ~ 2.00 instead of ~ 0.30
2. Many more postmasters are running ~ 20 instead of ~ 5

Some changes I made were:

1. Moved the log directories pg_xlog and pg_clog over to their own
drives (made symbolic links from there)

2. Changed the amount of shared memory from 350MB to 320MB

What would account for the new load? Queries seem to be taking much
longer. We have logs of RAM on the server 6GB and are running  SCSI Raid
on the system. This is also a Dual Xeon system (2.4GHz).

Some more background...
1. The database locked up last night (each postgres process was in
Uninterruptible Sleep)
2. We had to cycle the power on the server to get it going again.

Any help/insights would be greatly appreciated.

- Ericson Smith
eric@did-it.com
http://www.did-it.com




Re: Numerous postmaster processes after upgrading to 7.2.3

From
Ericson Smith
Date:
Ok, I found the problem.
Doing an analyze on the database solved it. Apparently the planner did
not have much to go on directly after the restore. Silly me.

- Ericson Smith
eric@did-it.com

On Mon, 2002-10-21 at 12:11, Ericson Smith wrote:
> Hi all,
>
> Today we upgraded to ver 7.2.3 from 7.2.2.
> I also dumped and restored the data before doing so, just in case :-)
>
> I've noticed a few things...
>
> 1. The database server load is much higher ~ 2.00 instead of ~ 0.30
> 2. Many more postmasters are running ~ 20 instead of ~ 5
>
> Some changes I made were:
>
> 1. Moved the log directories pg_xlog and pg_clog over to their own
> drives (made symbolic links from there)
>
> 2. Changed the amount of shared memory from 350MB to 320MB
>
> What would account for the new load? Queries seem to be taking much
> longer. We have logs of RAM on the server 6GB and are running  SCSI Raid
> on the system. This is also a Dual Xeon system (2.4GHz).
>
> Some more background...
> 1. The database locked up last night (each postgres process was in
> Uninterruptible Sleep)
> 2. We had to cycle the power on the server to get it going again.
>
> Any help/insights would be greatly appreciated.
>
> - Ericson Smith
> eric@did-it.com
> http://www.did-it.com
>
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org