- 04 Apr, 2017 35 commits
-
-
Brian Christiansen authored
END_TIMER was being called without a START_TIMER. Coverity CID:165991
-
Brian Christiansen authored
-
Isaac Hartung authored
-
Brian Christiansen authored
ACTIVE: accepts submissions and schedules INACTIVE: doesn't accept submissions or schedules DRAIN: schedules but doesn't accept submissions DRAIN+REMOVE: like DRAIN but cluster removes itself from federation once all federation jobs are gone.
-
Brian Christiansen authored
-
Brian Christiansen authored
-
Brian Christiansen authored
-
Brian Christiansen authored
Continuation of f9d132fc which just stopped the back_scheduler from starting up. This patch allows the backfill scheduler to be stopped after starting.
-
Isaac Hartung authored
Clearing should make the job eligible on all clusters again.
-
Isaac Hartung authored
-
Isaac Hartung authored
-
Brian Christiansen authored
cluster for it to handle the request.
-
Brian Christiansen authored
for ClusterFeatures.
-
Brian Christiansen authored
e.g. scontrol update <jobid> clusters=fed1,fed2
-
Brian Christiansen authored
-
Brian Christiansen authored
-
Brian Christiansen authored
Currently assumes that if the origin job updates successfully then it can submit the job to the siblings. Will need to expand on this to handle heterogenous setups.
-
Brian Christiansen authored
with the correct node select plugin_id. The dbd stores the plugins actual plugin id but the plugin_id doesn't index you into the plugin_context array. So it needs to be translated to the index so that the correct node_select plugin accessed properly.
-
Brian Christiansen authored
-
Isaac Hartung authored
-
Isaac Hartung authored
-
Isaac Hartung authored
-
Isaac Hartung authored
-
Isaac Hartung authored
-
Isaac Hartung authored
-
Isaac Hartung authored
If the pointer is NULL then the size_val should be zero or you'll be packing garbage.
-
Tim Wickberg authored
Compare the release locks to the acquired locks. More interestingly, in further commits this will allow us to annotate functions to verify the appropriate locks are held at runtime, and head off potential sources of corruption. Only enabled on development builds.
-
Tim Wickberg authored
Calling lock_slurmctld() while already holding locks leads to deadlock. Intentionally cause such instances to abort() on development builds to force these locations to the surface.
-
Tim Wickberg authored
-
Brian Christiansen authored
-
Brian Christiansen authored
when using fed_mgr_is_active(). It should only be used outside of any locks.
-
Tim Wickberg authored
Prior security check only permitted root or the SlurmUser to update.
-
Tim Wickberg authored
No way to do this through sbatch, but the API would permit it.
-
Thomas Opfer authored
Bug 3631.
-
Tim Wickberg authored
-
- 03 Apr, 2017 5 commits
-
-
Brian Christiansen authored
-
Brian Christiansen authored
-
Brian Christiansen authored
When not in a federation, or using --local, just go directly to the cluster without spawning a thread.
-
Brian Christiansen authored
-
Brian Christiansen authored
Instead of packing an empty federation rec, just pack a NULL and return a NULL.
-