Re: Immortal backend .... (kill -9 ?..) - Mailing list pgsql-admin

From Jeremy Buchmann
Subject Re: Immortal backend .... (kill -9 ?..)
Date
Msg-id 3CE3D62E.20409@wellsgaming.com
Whole thread Raw
In response to Immortal backend .... (kill -9 ?..)  ("Rajesh Kumar Mallah." <mallah@trade-india.com>(by way of Rajesh Kumar Mallah. <mallah@trade-india.com>))
List pgsql-admin
Rajesh Kumar Mallah. wrote:
> Hi,
>
> Two of the backends are showing status "D"  which "man ps"
> describes as "uninterruptible sleep (usually IO)"
>
> since kill -9 does not get me rid of these should i expect
> postmaster which is still the parent process for those two
> spoilt kids to terminate them?

That does sound like a hardware-related problem.  I saw the same thing
on a flaky Jaz drive.  The only thing I could do was reboot, which got
rid of the offending processes.  But that won't fix your hardware problem.

--Jeremy

> On Wednesday 15 May 2002 07:46 pm, Tom Lane wrote:
>
>>"Rajesh Kumar Mallah." <mallah@trade-india.com>(by way of Rajesh Kumar Mallah. <mallah@trade-india.com>) writes:
>>
>>>now even kill -9 <pid of backend> does not releases it below is
>>>the current status of backends:
>>
>>I think there is something broken about your system.  It is not normal
>>to be in a state where you have unkillable processes.  I'm wondering
>>about disk drive problems leading to I/O requests that never complete,
>>myself...
>>
>>            regards, tom lane
>
>




pgsql-admin by date:

Previous
From: "BELLIARD Francois - FT.BD/FTRD/DRHG/RSCL"
Date:
Subject: Re: problem in module Pg with Apache, Perl and SSL
Next
From: "Rajesh Kumar Mallah."
Date:
Subject: Help with data recovery...