Learn to run basic Pegasus examples on Verification Cluster
Use DM-8156 to properly allocate nodes and run jobs on the worker nodes.
Run prototype executor-less ci_hsc Pegasus workflow on Verification Cluster
add support for Slurm to allocateNodes.py
Greg Daues has installed Pegasus on the Verification Cluster:
I'm able to run some simple Pegasus workflow made in DM-7981 on Verification Cluster. I did not need DM-8156 to do so. There seem 4 condor slots available (and can be seen through condor_status) without doing ctrl_execute allocateNodes.py
I'll try a more complicated workflow in DM-8392.
Putting this back to "InProgress". I was running on local only, and not on Verification Cluster worker nodes.
We need to move pegasus to someplace more permanent. For condor, it's under
which is a link to
if we do the same for pegasus, we won't need to update that value in our git repos.
With DM-8156, I am able to run some simple workflows (testPegasus, testPegasusStack, exampleTask, basicProcessCcd here) on the Verification Cluster worker nodes using the given allocated node set.
For the Pegasus data staging configuration pegasus.data.configuration, I used condorio in one simplest workflow, and sharedfs for the other three.
I guess the current installation of Pegasus was meant to be temporary while in test. I'll leave the moving to Greg Daues.