[Gllug] kill -9 does not work

[Jake Jellinek]

Hi,

Most likely some form of fast memory leak or memory over commit and the
machine entered into "swap hell". It may have recovered hours and hours
later of it's own accord or it may not have. We often enjoy pressing the
"num lock" or "caps lock" keys to see if the lights respond when this
happens as a kind of indicator to how bad the problem is but doing that
remotely is not so easy ;-). The magic SysRq key can also help (in theory).

We do end up running a lot of monitoring scripts which automatically kill
processes that start to eat memory etc, as although Linux copes admirably
with resource over commitment, it nevertheless can't actually continue to
respond normally if software is allowed to eat up all available memory
and/or processor. It can seem odd sometimes that the remote shell continues
to respond to input apparently quite happily under these kinds of conditions
and the load average can still be apparently low, but the hanging and lack
of response, and the kill -9 having no apparent effect is something I've
seen a few times before. If you're lucky and you try hard you can eventually
kill something which is particularly memory intensive (a busy apache server
perhaps) and this can suddenly release the load enough to fully recover.

Not a very technical explanation or pure method of system administration I
know, but just a few random thoughts...

Jake.

----- Original Message -----
From: "Xander D Harkness" <xander at harkness.co.uk>
To: <gllug at linux.co.uk>
Sent: Tuesday, June 26, 2001 10:16 PM
Subject: Re: [Gllug] kill -9 does not work


> we are using level3 for some of our hosting (well racking and 30 feet of
> cable to our provider) and they have killed our connectivity  - dunno
> hacksawed through a cable or something - but anyway we have no external
> access to my lovely production boxes - so I took the opportunity to
errrrrm
> yes err reboot....
>
> I should mention also that the command ls would hang as well.
>
> All is new and well :-) my server is up and happy again.
>
> I would have like to know what the output and problem was.
>
> Cheers
> Xander
>
>
>
>
>
> Jake Jellinek wrote:
>
> > Hi,
> >
> > (assuming this is still running!)
> >
> > What is the process looking like in the process list if you do say a:-
> >
> > ps auxw | grep python
> >
> > ?
> >
> > Cheers,
> >
> > Jake.
> >
> > ----- Original Message -----
> > From: "Xander D Harkness" <xander at harkness.co.uk>
> > To: <gllug at linux.co.uk>
> > Sent: Tuesday, June 26, 2001 7:25 PM
> > Subject: [Gllug] kill -9 does not work
> >
> > > I have a production server where I want to upgrade rpm and other
files.
> > >
> > > I ran rpm -Uhv rpm-python*
> > >
> > > The command hung.  I tried to stop the process using Ctrl C
> > >
> > > The process is still running
> > >
> > > I tried killing the PID using kill and kill -9
> > >
> > > The process is still running
> > >
> > > I killed the bash and ssh session that initiated the command
> > >
> > > The process is still running
> > >
> > > Does anyone have any ideas (and please this would not be a suitable
> > > forum for apt-get blah, blah, blah ;-)
> > >
> > > Cheers
> > > Xander
> > >
> > >
> > >
> > >
> > > --
> > > Gllug mailing list  -  Gllug at linux.co.uk
> > > http://list.ftech.net/mailman/listinfo/gllug
> > >
> > >
> >
> > --
> > Gllug mailing list  -  Gllug at linux.co.uk
> > http://list.ftech.net/mailman/listinfo/gllug
>
>
> --
> Gllug mailing list  -  Gllug at linux.co.uk
> http://list.ftech.net/mailman/listinfo/gllug
>
>


-- 
Gllug mailing list  -  Gllug at linux.co.uk
http://list.ftech.net/mailman/listinfo/gllug