[lustre-discuss] Lustre OSS and clients on same physical server

Christopher J. Morrone morrone2 at llnl.gov
Wed Jul 13 12:00:34 PDT 2016

The bigger problem tends to be memory management issues in Lustre.  If
you put both the client and server code on the same node and do any
serious amount of IO, it has been pretty easy in the past to get that
node to go completely out to lunch thrashing on memory issues.  If your
server/client node is diskless, it will OOM and die.  If it has swap,
it'll swap itself into oblivion.  When lustre gets into that situation,
things start timing out and trying to reconnect, and then the situation
tends to cascade out of control and never recover from there without
rebooting nodes.

Which isn't to say that there aren't some workloads where you would be
just fine.  But memory management in lustre isn't good enough to
recommend running client and server on the same node for many applications.


On 07/13/2016 05:02 AM, Patrick Farrell wrote:
> It will work, but performance of both will suffer.  For that reason,
> it's usually not recommended.
> - Patrick
> ------------------------------------------------------------------------
> *From:* lustre-discuss <lustre-discuss-bounces at lists.lustre.org> on
> behalf of Amjad Syed <amjadcsu at gmail.com>
> *Sent:* Wednesday, July 13, 2016 4:05:30 AM
> *To:* lustre-discuss at lists.lustre.org
> *Subject:* [lustre-discuss] Lustre OSS and clients on same physical server
> Hello,
> I am new to lustre and currently doing a small POC for my  organization. 
> I would like to understand if Lustre OSS server and lustre client can be
> the same physical server  .
> If this setup can be used for production  ?
> Thanks
> _______________________________________________
> lustre-discuss mailing list
> lustre-discuss at lists.lustre.org
> http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org

More information about the lustre-discuss mailing list