Re: DB Restart - Mailing list pgsql-general

From Richard Huxton
Subject Re: DB Restart
Date
Msg-id 4AFC22AE.7030506@archonet.com
Whole thread Raw
In response to Re: DB Restart  (Sam Jas <samjas33@yahoo.com>)
List pgsql-general
Sam Jas wrote:
> We are running on 64 - bit.
> Whenever the reserved memory in top command it crosses 3.9g it hangs.
> If we try to kill process (using pg_cancel_backend()) it won't kill. At last either
> we have to kill all process at OS level or we have to reboot the server. After rebooting
> server everything worked perfectly fine. BTW we are using postgreSQL 8.3.2.
>
> PID       USER      PR
> NI  VIRT   RES
> SHR S %CPU %MEM    TIME+  COMMAND
>
>  4822 postgres  15   0
> 4045m 3.8g 3.8g S  0.7 12.1
> 2:09.63 postgres: writer process

Well, this process won't respond to pg_cancel_backend() because it's not
a normal backend. It's the background-writer process. Do you see the
same problem with ordinary backend processes, or is it always the writer?

Also - are you familiar with "strace"?

--
  Richard Huxton
  Archonet Ltd

pgsql-general by date:

Previous
From: Craig Ringer
Date:
Subject: Re: re-using RETURNING
Next
From: Tom Lane
Date:
Subject: Re: knowing which table/schema is going to be chosen