Fw: [Rhic-rcf-l] Jobs with memory leaks

mcbreen@sgs1.hirg.bnl.gov

--
Betty McBreen 631-344-5111 Fax 631-344-1334
----- Original Message ----- 
From: "RCF/USAtlas Staff" <rcfstaff@bnl.gov>
To: <rhic-rcf-l@lists.bnl.gov>; <rhic-software-l@lists.bnl.gov>
Sent: Friday, March 01, 2002 10:14 AM
Subject: [Rhic-rcf-l] Jobs with memory leaks


> As part of our effort to keep jobs with
> large memory leaks from crashing Linux
> nodes (we have had 2 such incidents this
> week and a few others in the recent past),
> we are implementing a script that will 
> terminate individual user processes if 
> it exceeds a pre-defined memory limit. 
> The default limit is 60% of the total
> available physical memory on a node.
> 
> The script will terminate both interactive
> and LSF jobs that exceed this limit. 
> 
> We will implement this on CAS nodes for
> now. We may implement it on CRS nodes in
> the future if the situation calls for it.
> 
> The implementation will begin at 12 noon 
> on Friday. If users object to this, please
> ask your RCF Liaison to contact me to make
> alternate arrangements. 
> 
> Tony
> 
> --
> This message forwarded from the RCF announcements page.
> Recent messages are available at:
> http://www.rhic.bnl.gov/RCF/Announcements/announce.html
> 
> _______________________________________________
> Rhic-rcf-l mailing list
> Rhic-rcf-l@lists.bnl.gov
> http://lists.bnl.gov/mailman/listinfo/rhic-rcf-l
>