Hi, I have changed SelectType from select/linear to select/cons_res and propagate slurm.conf and restart all slurmd and slurmctd. The next job I submitted was JobID 2 is that normal ? shouldn't it be $JOBID_BEFORE_RESTART + 1 ?
I am using slurmdbd configured like this : DbdHost=$SLURMDBD_HOST SlurmUser=slurm DebugLevel=3 AuthType=auth/none AuthInfo=/var/run/munge/munge.socket.2 LogFile=/var/log/slurm-llnl/slurmdbd.log PidFile=/var/run/slurm-llnl/slurmdbd.pid StorageType=accounting_storage/mysql StorageHost=localhost StorageUser=slurm StoragePass=$PASS and slurm.conf like this : ... SelectType=select/cons_res AccountingStorageType=accounting_storage/slurmdbd AccountingStorageHost=$SLURMDBD_HOST SlurmctldDebug=3 SlurmdDebug=3 ... But job information are stored and I have two entries in the DB for the same id_job : mysql> select job_db_inx,id_job from $CLUSTERNAME_job_table where id_job=63; +------------+--------+ | job_db_inx | id_job | +------------+--------+ | 82 | 63 | | 169 | 63 | +------------+--------+ 2 rows in set (0.00 sec) Is that normal ? Did I missed sthg ? Using slurm 2.2.1 on Debian 6 Regards, -- Stéphan
