Actions

Scyld Clusterware - User info

From Montana Tech High Performance Computing

Revision as of 10:05, 18 September 2017 by Bdeng (talk | contribs) (Created page with "Scyld ClusterWare information and commands available to general users will be added here. ==ClusterWare user level commands== ===bpstat=== Provides state, status, and statisti...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Scyld ClusterWare information and commands available to general users will be added here.

ClusterWare user level commands

bpstat

Provides state, status, and statistics for scheduling, monitoring, and diagnostics. See man pages for more info. To check processes running on all nodes:

ps aux | bpstat -P

On specific nodes:

ps aux | bpstat -P n1,n2

For a specific user:

ps aux | bpstat -P |grep username

For all process on management node, just use 'ps aux'

beostatus

Shows the status of the CPUs on the cluster. If not logged in with NX or running an X server, use:

beostatus -c

bpsh

Parallel shell access to one or more compute nodes. Some examples of how you might use it:

bpsh n0-n10 w
bpsh n0-n21 mkdir /scratch/username
bpsh n0,n3-n6 ls /scratch