cluster:162
This is an old revision of the document!
- Jobs can be submitted from any node.
cottontailis primary scheduler login node.- You can login in to any tail node directly (via ssh).
- All nodes are CentOS6.x with some exceptions noted.
Test Queue
- Walltime (CPULIMIT) has been removed (was 8 hrs/job)
- Jobslots raised to 48, is comprised of
greentailour /sancratch provider and redhat 5.5cottontail2physically identical togreentailfor fail over- standby scheduler master for
cottontail
swallowtailsecondary login nodepetaltail, sandbox and Warewolf provisioning noden29andn31, our OpenHPC test cluster compute nodes, CentOS 7.3
You can target individual node with '#BSUB -m node-name' in your submit script.
DR
sharptail2dris our disaster recovery node (to back upsharptail's /home)- Every couple of days users' /home will be refreshed on this DR node
- 17TB of usable storage
- Users selected for first or refresh backup are obtained from job submission logs last 30 days
- No logins are permitted
- Node will eventually be moved to Freeman DR location
- If users become inactive, backup home directory will not be removed (while space is available)
- vlan52 bonded interfaces (2G connection)
cluster/162.1504721673.txt.gz · Last modified: by hmeij07