Open broeder-j opened 6 years ago
Thanks @broeder-j for the report. First of all a note, I committed the LSF scheduler plugin but it was all work of @nmounet, who had the chance to test it only on one machine.
I see a few options to solve this (probably, in priority order according to me):
dateutil
package (it might be called python-dateutil
) is able to automatically detect both formats (it provides an automatic detection of format)If you can try if one of the above options could work (especially 1 and 2), it would be great (and in case we can discuss at the MaX Hackathon).
@broeder-j do you know if something like 1. above can be done? (i.e. if there is an ENVVAR that we can set to specify explicitly the time format?) I don't have a LSF cluster to test this.
On one of our machines running lsf time has the format '%b %d %H:%M:%S %Y ' which is not the expected format '%Y %b %d %H:%M L' by aiida....
Therefore, in the current implementation the lsf plugin won't work for this machine...
For our machine I have just hacked it. Through I do not know what other time formats might be out there and how we can in general cover for it... besides trying out a bunch of formats... From a quick check I have found also nothing in datetime that 'auto' recognizes the format, so I do not have a solution to push. Other schedulers plugins might have similar problems, I suspect...
Might be a thing for the coding week. Check how to parse in general different datetime strings. @giovannipizzi: Since you wrote it. Any suggestions?