Queued but does not start - Resources are not available (CPU)

Troubleshooting
1

Hello,
I just installed the cryosparc v2.1 and updated to 2.2. While running the tutorial with Empiar10025, the ‘Import movies’ work fine, but when executing Full frame motion correction, the job is queued but does not start. It indicates,
Lane default (node) 1
P1 J14
rigid_motion_correction 0
Waiting because resources are not available (CPU)

Tried many ways but could not troubleshoot, perhaps missing something silly during installation !!
This is running on a workstation with 64GB RAM and 1080 GTX (Cuda 9.1) Ubuntu 16.04, the standalone installation

Can someone provide me some clues?
Thanks,
T

2

Experiencing the same issue with Cryosparc 2.2. The scheduler sees I have only 4 cores, but the job has requested 6. This is a motion correction job, and I didn’t see any CPU count parameter in the job builder.

My command_core log:

Lane  default node : Jobs Queued (nonpaused, inputs ready):  [u'J11']
Total slots:  {u'somehost.somewhere': {u'GPU': set([0]), u'RAM': set([0, 1, 2, 3, 4, 5, 6, 7]), u'CPU': set([0, 1, 2, 3])}}
Available slots:  {u'somehost.somewhere': {u'GPU': set([0]), u'RAM': set([0, 1, 2, 3, 4, 5, 6, 7]), u'CPU': set([0, 1, 2, 3])}}
Available licen:  10000
Now trying to schedule J11
  Need slots :  {u'GPU': 1, u'RAM': 2, u'CPU': 6}
  Need fixed :  {u'SSD': False}
  Need licen :  True
  Master direct :  False
   Trying to schedule on somehost.somewhere
    Launchable:  False
    Alloc slots :  {}
    Alloc fixed :  {}
    Alloc licen :  False
    Queue status : waiting_resources
    Queue message : Waiting because resources are not available (CPU)
---------- Scheduler done ------------------
3

Thanks bmcgough,
Perhaps you figured the reason. I am also using a Xeon quadcore CPU. Not sure whether there is a way to override that need for 6 cpu.

Here is the relevant part of cryosparcm log command_core

---------- Scheduler running ---------------
Lane default node : Jobs Queued (nonpaused, inputs ready): [u’J16’]
Total slots: {u’tbcryoem’: {u’GPU’: set([0]), u’RAM’: set([0, 1, 2, 3, 4, 5, 6, 7]), u’CPU’: set([0, 1, 2, 3])}}
Available slots: {u’xxcryoem’: {u’GPU’: set([0]), u’RAM’: set([0, 1, 2, 3, 4, 5, 6, 7]), u’CPU’: set([0, 1, 2, 3])}}
Available licen: 10000
Now trying to schedule J16
Need slots : {u’GPU’: 1, u’RAM’: 2, u’CPU’: 6}
Need fixed : {u’SSD’: False}
Need licen : True
Master direct : False
Trying to schedule on xxcryoem
Launchable: False
Alloc slots : {}
Alloc fixed : {}
Alloc licen : False
Queue status : waiting_resources
Queue message : Waiting because resources are not available (CPU)

4

We tried changing the instance type to increase cpu count. Now with 8 cpus, cryosparc2 is still not running jobs and reports only 4 cpu slots. I tried deleting *.pyc files in hopes of eliminating any possible cached values, but that did not help. Is there a way to update the CPU count other than to re-install Cryosparc2?

5

Thanks for the info, and sorry I cannot answer your question. I am communicating with someone who is more knowledgeable but do not have an answer yet. I am also in process of acquiring a 8 core cpu (i.e. Xeon E5-2650v2 instead of E5 1607) for the same motherboard but have not tested it yet. Will post when I get to test it. It would have been nice if we could get in touch with one of the developers, especially if the requirements can be tweaked.