Hi Ron, I also am in the "paranoid" group. And I've always done updates with jobs "live". Depending on the size of your userbase you may want to consider pausing the submission/start of new jobs while you execute the dnf commands (yes, I use them, rather than the "raw" rpm, because I think they are less error prone with e.g. dependencies). Since you are in the same group as myself, you can save a list of running jobs before and after executing the dnf commands, and see if they match. If they do, congratulations, everything went well. If they don't, there is a (tiny) risk that the jobs which completed during that time might miss "something". Examine their logs and/or warn the users as appropriate. To be clear, this tiny risk is about jobs that would complete *on their own* during that timeframe, not that the slurm update will cause healy jobs to crash. What could happen is a race condition between the jobs terminating and the slurm update which might try to update some information in some DB in an inconsistent way. My understanding is that the job itself (e.g. output file) are safe, it's just the slurm records which might get some trouble. You mention "waiting at least a week" between a subsequent update, but really the key point is this *Before considering the upgrade complete, **wait for all jobs that were already running to finish**.* Which means: if you have a 6h wallclock limit, you can wait only 7h. If you have a 2 months wallclock limit you need to wait for a bit more than 2 months. If you don't have wallclock limit.... you may have to wait forever.... Wait! You have the list of jobs because you are paranoid like myself and made one as mentioned above, so you have to wait "only" for all of them to be completed before proceeding, not "forever". With these precautions, most likely you won't encounter any issue (of course that gets weighted with the size of the cluster: if you have a huge one with hundreds of thousands of users/jobs/nodes, you will see things that have 0.001% chance of happening and that most of us never encounter) HTH. On Tue, Jan 20, 2026 at 1:31 PM Ron Gould via slurm-users < slurm-users@lists.schedmd.com> wrote:

Thank you for that guidance. I am certainly in the "overly cautious" and "paranoid" groups.

I will probably go through the slower upgrade process (1-8 list), with at least a week between them.

And yes, if anyone has experience doing such a vault between versions, please chime in.

-- slurm-users mailing list -- slurm-users@lists.schedmd.com To unsubscribe send an email to slurm-users-leave@lists.schedmd.com

Hi Ron, On 1/20/26 22:53, Ron Gould via slurm-users wrote:

Thank you for your pointers and sharing your experience.

We always upgrade Slurm while the cluster (700 nodes) is running production jobs, and we never had any issues. As Davide said, the chance of errors seems to be very small. Minor version upgrades should be simple to do because Slurm is basically unchanged. Major version upgrades should be done a little more carefully, just to be on the safe side. I have collected information on Slurm upgrading, database dumps etc. in these Wiki pages: https://wiki.fysik.dtu.dk/Niflheim_system/Slurm_installation/#upgrading-slur... https://wiki.fysik.dtu.dk/Niflheim_system/Slurm_database/#backup-and-restore... https://wiki.fysik.dtu.dk/Niflheim_system/Slurm_database/#backup-and-restore... Please beware of a MariaDB upgrade issue that was resolved in 22.05.7: https://wiki.fysik.dtu.dk/Niflheim_system/Slurm_database/#slurm-database-mod... IHTH, Ole

My user base is likely small compared to other institutions. Currently, I have about 10 users running about 30 jobs, with some started today and the oldest started in September.

Regarding the "waiting a week" between updates, most of the jobs are short lived, with some taking less than a week. Given that I don't have a short WallClock value, I could update to 23.11 before those long jobs would have to be stopped and restarted under the new slurm dæmons. Doing a couple updates would give me ample practice and I can document the entire thing.

My "slurm_acct_db" database, I have daily, weekly, and monthly backups of it. It's under 2 GB if I had to re-import it. I don't suspect the slurmdbd upgrade will take long.

Prior to that DB backup, I have another script that backs up `${StateSaveLocation}` and "/etc/slurm". This is referenced in "https://slurm.schedmd.com/upgrades.html#backups>>

-- Ole Holm Nielsen PhD, Senior HPC Officer Department of Physics, Technical University of Denmark

Whoops, no I am using MariaDB.

Thank you for that point. That had occurred to me, but as this is my first upgrade, I just wanted to upgrade to the version that fixes the bug my users see. I need a li'l win for this Cluster :) . When I have that, I'll take a couple more baby steps and do 22.05.11 with a quick turnaround.

Similar to you we host our slurmdbd and slurmctld on the same host. So when we upgrade we follow the following steps: 1. Stop slurmctld 2. Stop slurmdbd 3. Take backups of the slurm database and slurm spool directory 4. Update rpms (make sure to neutered the autorestart function in the rpm spec as that can restart things before you are ready) 5. Run slurmdbd -D in a tmux (as the upgrade can take a while and if you use systemd it will terminate early). 6. Once slurmdbd ugprade is complete stop the command line version of slurmdbd and restart it using systemd 7. Global restart of slurmctld and slurmd for the cluster. -Paul Edmon- On 2/3/26 9:20 AM, Ron Gould via slurm-users wrote:

Hi there.

The SLURM DB service is installed alongside the other SLURM packages on the node I'm starting with.

slurm-slurmdbd-22.05.3-1.el7.x86_64 slurm-22.05.3-1.el7.x86_64 slurm-contribs-22.05.3-1.el7.x86_64 slurm-devel-22.05.3-1.el7.x86_64 slurm-example-configs-22.05.3-1.el7.x86_64 slurm-libpmi-22.05.3-1.el7.x86_64 slurm-openlava-22.05.3-1.el7.x86_64 slurm-pam_slurm-22.05.3-1.el7.x86_64 slurm-perlapi-22.05.3-1.el7.x86_64 slurm-slurmctld-22.05.3-1.el7.x86_64 slurm-slurmd-22.05.3-1.el7.x86_64 slurm-torque-22.05.3-1.el7.x86_64

It doesn't seem like I am able to upgrade that one package first like the instructions say to. Is it okay to upgrade all of the SLURM packages in one go? If so, I would stop the slurmdbd service, back up the appropriate directories, back up the DB into a self-contained file, upgrade the packages all together, start the slurmdbd service ( `sudo -u slurm slurmdbd -D` ), and so forth according to the instructions.

Is that a cogent plan?

Yeah, we use dnf. I do: dnf update slurm slurm-libpmi slurm-devel slurm-contribs slurm-slurmctld slurm-perlapi slurm-slurmrestd slurm-slurmdbd For the autorestart compare: https://github.com/SchedMD/slurm/blob/cead8b9d2c2360f976c77d9e9e7ab875de9d86... With: https://github.com/fasrc/slurm-spec/blob/75ec489f2a8da057432495c0a726ed80855... You can see that in the official spec it has: %systemd_postun_with_restart slurmdbd.service Which means when the rpm is uninstalled it restarts the service. You don't want that to happen in this case as that can mess up the slurmdbd (unless you increase the systemd timeouts). So in our spec we nerf it and understand that we have to force a restart rather than letting dnf do the restart itself. Basically its a paranoia step we take as we have been bitten by it before as our database got messed up due to this exact thing happening and we had to reimport our backup (which is why you take a backup) -Paul Edmon- On 2/3/26 10:51 AM, Ron Gould via slurm-users wrote:

Thank you Paul. That helps a lot.

Regarding item 4., do you use `dnf` or `rpm` to install the RPMs? What's your syntax?

Also in 4., I may need to rebuild my RPMs anway, as I may have to update an option. What's the option to "neutered the autorestart function"?

Thanks for your help, Ron

We generally maintain our own spec for that reason. It's just a clone of the spec that Slurm provides but with our edits. That repo I linked is what we use for maintaining that spec with a branch for each version (https://github.com/fasrc/slurm-spec) -Paul Edmon- On 2/3/26 12:05 PM, Ron Gould via slurm-users wrote:

Interesting. Is that something I can do with the existing code as downloaded (TAR.XZ file) or do I need to fork it and make changes?

Well, I have the code downloaded. I can extract it and modify it locally.

Maybe I've answered my own question. Please confirm.

That's because your previous version had the restart in it. The restart unfortunately happens when the package is uninstalled, and thus applies to the previous version, not the new version. There isn't much you can do about that you will just need to be careful. -Paul Edmon- On 2/4/26 1:08 PM, Ron Gould via slurm-users wrote:

I modified the "slurm.spec" file in the TAR.BZ2 file by prepending "###" to the

`%systemd_postun_with_restart slurmdbd.service`

line at the end, in the "%postun" section and added

`%systemd_postun slurmdbd.service`

That didn't seem to work.

I still ended up with a postuninstall script that tries to restart the slurmdbd service:

``` # rpm -q --scripts "rpmbuild.20260204/RPMS/x86_64/slurm-slurmdbd-22.05.4-1.el8.x86_64.rpm" postinstall scriptlet (using /bin/sh):

if [ $1 -eq 1 ] ; then # Initial installation systemctl --no-reload preset slurmdbd.service &>/dev/null || : fi preuninstall scriptlet (using /bin/sh):

if [ $1 -eq 0 ] ; then # Package removal, not upgrade systemctl --no-reload disable --now slurmdbd.service &>/dev/null || : fi postuninstall scriptlet (using /bin/sh): ### if [ $1 -ge 1 ] ; then # Package upgrade, not uninstall systemctl try-restart slurmdbd.service &>/dev/null || : fi ```

The prepended "###" on that line might've been the problem. If I just changed the SPEC file to

``` ... %post slurmdbd %systemd_post slurmdbd.service %preun slurmdbd %systemd_preun slurmdbd.service %postun slurmdbd %systemd_postun slurmdbd.service ``` ,

I then get a better looking postuninstall script:

``` # rpm -q --scripts "rpmbuild.20260204 Modified SPEC v1/RPMS/x86_64/slurm-slurmdbd-22.05.4-1.el8.x86_64.rpm" postinstall scriptlet (using /bin/sh):

if [ $1 -eq 1 ] ; then # Initial installation systemctl --no-reload preset slurmdbd.service &>/dev/null || : fi preuninstall scriptlet (using /bin/sh):

if [ $1 -eq 0 ] ; then # Package removal, not upgrade systemctl --no-reload disable --now slurmdbd.service &>/dev/null || : fi postuninstall program: /bin/sh ```

Is having an empty "postuninstall program: /bin/sh" entry okay?

Probably. I've never tried that. I imagine that will work. -Paul Edmon- On 2/5/26 2:56 AM, Steffen Grunewald wrote:

Stupid (?) question: wouldn't a `systemctl disable slurmdbd.service` be the best choice? I'm a Debian user and in the past this strategy seems to have worked, with the packages provided by Debian developers/maintainers ... is RPM really that different?

Thanks, S

On Wed, 2026-02-04 at 13:27:10 -0500, Slurm users wrote:

...

That's because your previous version had the restart in it. The restart unfortunately happens when the package is uninstalled, and thus applies to the previous version, not the new version. There isn't much you can do about that you will just need to be careful.

-Paul Edmon-

On 2/4/26 1:08 PM, Ron Gould via slurm-users wrote:

...

I modified the "slurm.spec" file in the TAR.BZ2 file by prepending "###" to the

`%systemd_postun_with_restart slurmdbd.service`

line at the end, in the "%postun" section and added

`%systemd_postun slurmdbd.service`

That didn't seem to work.

Good to know. I'll add that to the documentation I made.