Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.
Bug 1947878
Summary: | Request to package Slurmrestd for EPEL7 and EPEL8 | ||
---|---|---|---|
Product: | [Fedora] Fedora EPEL | Reporter: | Heitor <heitorpbittencourt> |
Component: | slurm | Assignee: | Philip Kovacs <pkfed> |
Status: | CLOSED ERRATA | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | unspecified | Docs Contact: | |
Priority: | unspecified | ||
Version: | epel7 | CC: | peter.kandolf, pkfed |
Target Milestone: | --- | Keywords: | Reopened |
Target Release: | --- | ||
Hardware: | Unspecified | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | slurm-20.11.6-1.el8 slurm-20.11.6-1.el7 slurm-20.11.7-3.el8 slurm-20.11.7-3.el7 | Doc Type: | If docs needed, set a value |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2021-06-09 02:14:13 UTC | Type: | Bug |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Bug Depends On: | 1940198 | ||
Bug Blocks: |
Description
Heitor
2021-04-09 13:18:14 UTC
The dependencies are all met for Fedora rawhide, so I'll implement this first in rawhide (which is also in the spirit of what rawhide is all about). I notice that slurm upstream mentions that json web tokens is optional for slurmrestd -- it would be heavily crippled without it, correct? libjwt is used for authentication, using json web tokens. Auth is optional, but in practice we always want authentication. I've added the slurmrestd sub-package to Fedora rawhide, so there is a testing ground for this now. Adding the libjwt bug as a blocker for this so the bottleneck is clear. I'm interested in this as well! If there is anything I can do to help and speed up the process I'm happy to. The libjwt bug (below) is on QA in the 2-week testing period with 6 days left to stable. After that goes through, I can submit the needed slurm changes which in turn will require another 2 week testing period. Testers can grant the builds karma to speed up the process. https://bugzilla.redhat.com/show_bug.cgi?id=1940198 https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-b8c4714317 https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-f1920cd4c1 FEDORA-EPEL-2021-f06db8686b has been submitted as an update to Fedora EPEL 8. https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-f06db8686b FEDORA-EPEL-2021-77115f5c43 has been submitted as an update to Fedora EPEL 7. https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-77115f5c43 FEDORA-EPEL-2021-f06db8686b has been pushed to the Fedora EPEL 8 testing repository. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-f06db8686b See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates. FEDORA-EPEL-2021-77115f5c43 has been pushed to the Fedora EPEL 7 testing repository. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-77115f5c43 See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates. Works perfect. My login https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-77115f5c43 is not working (500 server error) so I give my feedback here and hope for good karma :) FEDORA-EPEL-2021-f06db8686b has been pushed to the Fedora EPEL 8 stable repository. If problem still persists, please make note of it in this bug report. FEDORA-EPEL-2021-77115f5c43 has been pushed to the Fedora EPEL 7 stable repository. If problem still persists, please make note of it in this bug report. I tried to setup a simple cluster with 4 nodes, one for each slurm component: slurmd, slurmctld, slurmdbd, and slurmrestd. Unfortunately, slurm{d,dbd,ctld} could not start, as they do not have the auth_jwt plugin (/usr/lib64/slurm/auth_jwt.so). Here is the end of the logs: [2021-05-24T16:58:27.330] debug3: Trying to load plugin /usr/lib64/slurm/auth_jwt.so [2021-05-24T16:58:27.330] debug4: /usr/lib64/slurm/auth_jwt.so: Does not exist or not a regular file. [2021-05-24T16:58:27.330] error: Couldn't find the specified plugin name for auth/jwt looking at all files [2021-05-24T16:58:27.330] error: cannot find auth plugin for auth/jwt [2021-05-24T16:58:27.330] error: cannot create auth context for auth/jwt [2021-05-24T16:58:27.330] fatal: failed to initialize authentication plugin O copied the file /usr/lib64/slurm/auth_jwt.so from slurmrestd node, installed libjwt, and restarted the slurm{d,dbd,ctld} services and the cluster came back to life. I think this file (/usr/lib64/slurm/auth_jwt.so) should be packaged in slurm-slurmdbd, slurm-slurmctld, and slurm-slurmdbd. Should I open another ticket for that? We'll use this ticket since the bug manifests from this new packaging -- this is why cluster testing is important! I agree that auth_jwt.so probably needs to be moved, but it must sit in one package only. Let me investigate further. Reopened. FEDORA-EPEL-2021-0e0c1a76c6 has been submitted as an update to Fedora EPEL 8. https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-0e0c1a76c6 FEDORA-EPEL-2021-7e9a7ecfb4 has been submitted as an update to Fedora EPEL 7. https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-7e9a7ecfb4 I moved the auth_jwt.so plugin into the base slurm package since it is needed for all node types (compute node, control node, database accounting node, etc.). This should resolve the problem. The other rest related plugins (openapi, rest_auth_jwt) are loaded only by the slurmrestd daemon itself, so those can stay in the slurmrestd subpackge. Thank you! Do you know when this fix will arrive in Epel7 stable? There's a 14-day time gate or +3 karma gate, whichever happens first. As soon as the update is swept to the testing repo (should be within 24 hours), however, you can use the command at the top of the update to test it. Essentially you activate the testing repo and then install that " advisory" package. Please test it if you can instead of waiting the 2 weeks, finding another problem, and then having to wait all over again. https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-7e9a7ecfb4 Once the package goes stable you can remove the advisory package installation, de-activate the testing repo and install it through the stable channel. FEDORA-EPEL-2021-0e0c1a76c6 has been pushed to the Fedora EPEL 8 testing repository. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-0e0c1a76c6 See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates. FEDORA-EPEL-2021-7e9a7ecfb4 has been pushed to the Fedora EPEL 7 testing repository. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2021-7e9a7ecfb4 See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates. FEDORA-EPEL-2021-0e0c1a76c6 has been pushed to the Fedora EPEL 8 stable repository. If problem still persists, please make note of it in this bug report. FEDORA-EPEL-2021-7e9a7ecfb4 has been pushed to the Fedora EPEL 7 stable repository. If problem still persists, please make note of it in this bug report. I can confirm slurmrestd is working nicely :) Thanks for packaging it, Philip Kovacs! |