Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[farmshare-discuss] FarmShare maintenance complete

Adam Seishas aseishas at stanford.edu
Sun Sep 24 09:34:20 PDT 2017


It looks like /farmshare/user_data wasn’t being mounted on corn02 and corn12; that should be fixed now.

—
Adam Seishas
Research Computing Specialist
Stanford | University IT

Sent from my iPad

On Sep 24, 2017, at 9:06 AM, Haben Michael <haben.michael at stanford.edu<mailto:haben.michael at stanford.edu>> wrote:

Nevermind, I see it's not that the issue came and went, it's that it only affects some corn machines.

On Sun, Sep 24, 2017 at 11:10 AM, Haben Michael <haben.michael at stanford.edu<mailto:haben.michael at stanford.edu>> wrote:
Hi Ling, Did you hear back about this? I briefly had this issue last week, then things were working fine through yesterday, then the problem returned this morning. farmshare/user_data is a symlink to nothing. There are some other kinks also, like both new and old farmshare are having difficulties with gridengine.

On Fri, Sep 22, 2017 at 7:09 PM, Ling Liu <lingliu at stanford.edu<mailto:lingliu at stanford.edu>> wrote:
Hi Adam,

Every since the power outage, I haven't been able to cd into the user_data directory in Farmshare and yet it works in Farmshare2. It kept saying "no such directory" while I can see it in the parental directory. Am I the only person having this problem? Thanks!

Ling

On Thu, Sep 21, 2017 at 5:36 PM, Adam Seishas <aseishas at stanford.edu<mailto:aseishas at stanford.edu>> wrote:
Both the FarmShare 2 and legacy FarmShare environments have been returned to service. FarmShare 2 service is slightly degraded pending additional maintenance on hardware affected by yesterday’s power outage.

During maintenance some new limits on FarmShare 2 compute jobs were introduced. You must now request a minimum of 96GB of memory (--mem=96G) when submitting a job to the bigmem queue; this change makes FarmShare submission requirements more similar to those on Sherlock 2.0. Limits have also been placed on the number of jobs that may be queued per user: 1024 on the normal partition, and 64 each on the long, bigmem, and gpu partitions. In order to encourage proper use of the interactive quality-of service, jobs submitted to this QoS will receive a modest priority boost. Finally, some additional limits will be introduced in the next few days that should give pending jobs more opportunities to run when there is contention for cluster resources.

—
Adam Seishas
Research Computing Specialist
Stanford | University IT
aseishas at stanford.edu<mailto:aseishas at stanford.edu> · 650.725.7490<tel:6507257490>





_______________________________________________
farmshare-discuss mailing list
farmshare-discuss at lists.stanford.edu<mailto:farmshare-discuss at lists.stanford.edu>
https://mailman.stanford.edu/mailman/listinfo/farmshare-discuss


_______________________________________________
farmshare-discuss mailing list
farmshare-discuss at lists.stanford.edu<mailto:farmshare-discuss at lists.stanford.edu>
https://mailman.stanford.edu/mailman/listinfo/farmshare-discuss


_______________________________________________
farmshare-discuss mailing list
farmshare-discuss at lists.stanford.edu<mailto:farmshare-discuss at lists.stanford.edu>
https://mailman.stanford.edu/mailman/listinfo/farmshare-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/farmshare-discuss/attachments/20170924/645ca84b/attachment.html>


More information about the farmshare-discuss mailing list