15.10 does checking that the job script written can be executed before
the job is submitted and some file system syncing that really may help
this issue. Hopefully this helps this particular case, in general
though Galaxy does need to be better about resuming workflows over
collections and retrying
Cluster has workers, jobs running on main node is disabled.
2015-08-03 14:44 GMT-05:00 John Chilton :
> Are you running jobs on the head node or just Galaxy? If this is a
> consistent problem and you are running jobs on the head ndoe I would
> disable that.
>
> As to resume just the failed jobs -
Are you running jobs on the head node or just Galaxy? If this is a
consistent problem and you are running jobs on the head ndoe I would
disable that.
As to resume just the failed jobs - this is not currently possible but
really should be ideally.
https://trello.com/c/lxVJy7fs
-John
On Mon, Jul