[slurm-users] SlurmDBD 20.02.7

Danny Marc Rotscher danny.rotscher at tu-dresden.de
Fri Jan 7 05:34:35 UTC 2022


Hi Bas,

thank you very much for linking the bug report, the following solution mentioned there helps me to solve the problem.

1. Stop slurmdbd
systemctl stop slurmdbd

We running SlurmDBD in Docker with Supervisord, so we have to stop it like follows.
supervisorctl stop slurmdbd

2. Modify the default values
Choose your MySQL/MariaDB client to access the database and modify the DEFAULT values from there.
use slurm_acct_db;
ALTER TABLE taurus_assoc_table MODIFY max_tres_pj text DEFAULT ''; 
ALTER TABLE taurus_assoc_table MODIFY max_tres_pn text DEFAULT '';
ALTER TABLE taurus_assoc_table MODIFY max_tres_mins_pj text DEFAULT '';
ALTER TABLE taurus_assoc_table MODIFY max_tres_run_mins text DEFAULT '';
ALTER TABLE taurus_assoc_table MODIFY grp_tres text DEFAULT '';
ALTER TABLE taurus_assoc_table MODIFY grp_tres_mins text DEFAULT '';
ALTER TABLE taurus_assoc_table MODIFY grp_tres_run_mins text DEFAULT '';
ALTER TABLE taurus_assoc_table MODIFY qos blob DEFAULT '';
ALTER TABLE taurus_assoc_table MODIFY delta_qos blob DEFAULT ‚';

Replace „taurus" with your cluster name.
 
3. Start slurmdbd again
systemctl start slurmdbd

Docker+Supervisord:
supervisorctl start slurmdbd

Kind regards,
Danny Rotscher


> Am 06.01.2022 um 16:26 schrieb Bas van der Vlies <bas.vandervlies at surf.nl>:
> 
> Hi Danny,
> 
> We had the same issue when we upgraded slurm to 20.11 but maybe the solution also works for you:
> * https://bugs.schedmd.com/show_bug.cgi?id=12947
> 
> On 06/01/2022 15:36, Danny Marc Rotscher wrote:
>> Hello everyone,
>> today we update our Slurm database daemon from 20.02.2 to 20.02.7 and everything works except when I try to delete a user it does not work.
>> sacctmgr -i delete user name=xyz account=xyz
>> sacctmgr: slurmdbd: No error
>>  Nothing deleted
>> slurmdbd.log:
>> slurmdbd_1  | slurmdbd: error: mysql_query failed: 1364 Field 'max_tres_pj' doesn't have a default value
>> slurmdbd_1  | update "taurus_assoc_table" as t1 set mod_time=1641479365, deleted=1, def_qos_id=DEFAULT, shares=DEFAULT, max_jobs=DEFAULT, max_jobs_accrue=DEFAULT, min_prio_thresh=DEFAULT, max_submit_jobs=DEFAULT, max_wall_pj=DEFAULT, max_tres_pj=DEFAULT, max_tres_pn=DEFAULT, max_tres_mins_pj=DEFAULT, max_tres_run_mins=DEFAULT, grp_jobs=DEFAULT, grp_submit_jobs=DEFAULT, grp_jobs_accrue=DEFAULT, grp_wall=DEFAULT, grp_tres=DEFAULT, grp_tres_mins=DEFAULT, grp_tres_run_mins=DEFAULT, qos=DEFAULT, delta_qos=DEFAULT, priority=DEFAULT where (id_assoc=3527);
>> slurmdbd_1  | slurmdbd: error: CONN:9 No error
>> Does anyone knows what that mean?
>> I also found out, that all *tres* fields in taurus_assoc_table are empty but not NULL.
>> Kind regards,
>> Danny Rotscher
> 
> -- 
> Bas van der Vlies
> | HPCV Supercomputing | Internal Services  | SURF | https://userinfo.surfsara.nl |
> | Science Park 140 | 1098 XG Amsterdam | Phone: +31208001300 |
> |  bas.vandervlies at surf.nl

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5092 bytes
Desc: not available
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20220107/084c42bc/attachment.bin>


More information about the slurm-users mailing list