[LON-CAPA-admin] Runaway memory?

[Mark Lucas]

Guy,

I haven't really had a chance to identify if it's just one process or
the collection.

I'm looking at capa8 right now and I have three children above 180Mb
(183-221). The rest are around 29-48.

Could an upload of a large file (say a technical report at 20Mb)
permanently bloat an httpd child?

I know I could tune the number of hits before a child is reloaded. Is
there a way to trim a child out when it gets too big (say over 100Mb)?

Later,
Mark


On Wed, 2007-10-24 at 14:34 -0400, Guy Albertelli II wrote:
> Hi Mark,
> 
> > Has anyone had any runaway memory events on their access servers?
> 
> Is it individual httpds becoming huge? Or just a large number of
> httpds?
> 
> Or is it some other process becoming huge?
> 
> > 
> > The out of memory killer killed my processes running on one of our
> > access servers during a moderately high load time. I'm running a
> > reasonably up to date CENTOS 5 version (the latest kernel update has not
> > yet been installed).
> > 
> > This may be concurrent with students trying to upload large files for a
> > lab report.
> > 
> > The memory started spiking about an hour before the crash. I began
> > getting lowmem_reserve errors on the console.
> > 
> > Any advice is welcome!
> > 
> > Mark
> > 
> > -- 
> > ----------------------------------------------------------------------------
> > Mark Lucas                                      email: lucasm at ohiou.edu
> > 252D Clippinger Lab                             phone: (740)597-2984
> > Department of Physics and Astronomy             fax:   (740)593-0433
> > Ohio University
> > Athens, OH 45701
> > 
> > _______________________________________________
> > LON-CAPA-admin mailing list
> > LON-CAPA-admin at mail.lon-capa.org
> > http://mail.lon-capa.org/mailman/listinfo/lon-capa-admin
>