[lustre-discuss] HPC Head node clustering and lustre

Brian Andrus toomuchit at gmail.com
Tue Nov 21 13:05:51 PST 2017


That would depend on how they setup PCS. If you are going to have a 
'floating' IP that is failed over and merely used for the login traffic, 
there should be no issue.

I think the method for HA is a bit overkill, however. Depending on what 
services are running, it may make more sense to just have a secondary 
system that is running and use DNS to failover or just round-robin users 
to balance them across the systems that are up. Certainly easier to 
maintain and easier to recover when something happens.

Brian Andrus

On 11/21/2017 4:44 AM, Amjad Syed wrote:
> Hello,
> We have a small 10 node compute cluster . We have single 
> management/head node which is used as login node as well. The 
> management consider this head node as single point of failure.  They 
> are planning to buy another head/login/management node and make an 
> active-passive cluster using Red hat pcs. The question i have is the 
> Lustre file system which is mounted on these two nodes , will it be 
> effected when the transfer between two nodes happen.
> _______________________________________________
> lustre-discuss mailing list
> lustre-discuss at lists.lustre.org
> http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20171121/498b7017/attachment-0001.html>

More information about the lustre-discuss mailing list