You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Bert DeKnuydt provided reported an issue as follows:
But 'condor_watch_q -user ' will error out
for non-priviledged users as follows:
(And often for priviledged users too, if the logfiles are over
e.g. NFS)
~ » id
uid=4653(deknuydt) gid=5800(visics) groups=5800(visics),1207(vdi-users),6200(xvisics)
~ » id xma
uid=35402(xma) gid=5800(visics) groups=5800(visics),1207(vdi-users)
~ » condor_watch_q -user xma
WARNING: Could not open event log at /esat/vauxite/xma/sing_images/ch-resnet-cifar10/scripts/condor/condor_files/2020-09-30-01-58-vgg11-adam-0.01-0.089spjpe9/condor_job.log for reading, so it will be ignored. Reason: JobEventLog not initialized. Check the debug log, looking for ReadUserLog or FileModifiedTrigger. (Or call htcondor.enable_debug() and try again.)
WARNING: Could not open event log at /esat/vauxite/xma/sing_images/ch-resnet-cifar10/scripts/condor/condor_files/2020-09-30-01-58-resnet8-sgd-0.1-0.001btxpvurh/condor_job.log for reading, so it will be ignored. Reason: JobEventLog not initialized. Check the debug log, looking for ReadUserLog or FileModifiedTrigger. (Or call htcondor.enable_debug() and try again.)
Processing new events...ERROR: Unhandled error: list index out of range. Re-run with -debug for a full stack trace.
The reason is simply that the logfiles are not readable (permission denied).
The above would make you think something is seriously wrong, but it isn't
You could easily clarify the error a bit, but that does not solve the real
problem of permissions; probably not fixable at all. Or do you see
a workaround?
The text was updated successfully, but these errors were encountered:
If HTCondor isn't doing a good job providing error messages about unreadable files, then we should simply test-open the file first using a normal python open.
Whatever is causing this issue (ERROR: Unhandled error: list index out of range) likely needs to be fixed. Maybe something is assuming the list of usable files is non-empty?
Bert DeKnuydt provided reported an issue as follows:
But 'condor_watch_q -user ' will error out
for non-priviledged users as follows:
(And often for priviledged users too, if the logfiles are over
e.g. NFS)
~ » id
uid=4653(deknuydt) gid=5800(visics) groups=5800(visics),1207(vdi-users),6200(xvisics)
~ » id xma
uid=35402(xma) gid=5800(visics) groups=5800(visics),1207(vdi-users)
~ » condor_watch_q -user xma
WARNING: Could not open event log at /esat/vauxite/xma/sing_images/ch-resnet-cifar10/scripts/condor/condor_files/2020-09-30-01-58-vgg11-adam-0.01-0.089spjpe9/condor_job.log for reading, so it will be ignored. Reason: JobEventLog not initialized. Check the debug log, looking for ReadUserLog or FileModifiedTrigger. (Or call htcondor.enable_debug() and try again.)
WARNING: Could not open event log at /esat/vauxite/xma/sing_images/ch-resnet-cifar10/scripts/condor/condor_files/2020-09-30-01-58-resnet8-sgd-0.1-0.001btxpvurh/condor_job.log for reading, so it will be ignored. Reason: JobEventLog not initialized. Check the debug log, looking for ReadUserLog or FileModifiedTrigger. (Or call htcondor.enable_debug() and try again.)
Processing new events...ERROR: Unhandled error: list index out of range. Re-run with -debug for a full stack trace.
The reason is simply that the logfiles are not readable (permission denied).
You could easily clarify the error a bit, but that does not solve the real
problem of permissions; probably not fixable at all. Or do you see
a workaround?
The text was updated successfully, but these errors were encountered: