Manager integration stability and correctness [Part 2]

[rzetelskik]

Part 2 of #1897

Task list

Preview Give feedback

1. Scylla Manager controller will update tasks despite no changes in spec #1827
   kind/bug priority/important-soon triage/accepted +3
   
2. Manager controller recreates clusters when manager cluster ID is missing from status #1902
   kind/bug priority/important-soon triage/accepted +3
   
3. Cover configuring CQL credentials in Scylla Manager integration docs pages #1896
   kind/feature priority/important-soon triage/accepted +3
4. Scylla Manager controller should remove clusters from manager state on object deletion #1930
   kind/feature priority/important-longterm triage/accepted +3
5. Add status conditions to manager controller #1921
   kind/feature priority/backlog triage/accepted +3
6. Refactor docs related to Scylla Manager integration #1898
   kind/documentation priority/important-longterm triage/accepted +3
7. Reevaluate non-optional or defaulted fields in manager task spec #1869
   kind/feature priority/important-longterm triage/accepted +3
8. Add missing fields to manager task status #1904
   kind/feature priority/backlog triage/accepted +3
9. Add an e2e suite covering manager API conversion to test unhandled properties #1870
   kind/feature priority/backlog triage/accepted +3
10. ScyllaClusters connected into a multi-datacenter cluster with externalSeeds are registered as separate clusters with Scylla Manager #2119
    kind/bug priority/important-longterm triage/accepted +3
11. Manager controller should wait until ScyllaCluster is Available before reconcilation
12. Document configuring workload identities with Scylla Manager Agent #1697
    kind/documentation priority/important-longterm triage/accepted +3

/kind epic
/assign

[rzetelskik]

#1827 and #1902 are kind of high priority, but we're waiting for a feature from the manager team. The rest won't make it for 1.13 imo.

[rzetelskik]

@tnozicka I also added a couple of issues that I created in the meantime, please verify if all fit here

[tnozicka]

np, I take it as a rolling one. We'll do the same dance at the end of the next release and make Part 3, if needed.

[scylla-operator-bot]

The Scylla Operator project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 30d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out

/lifecycle stale

[rzetelskik]

/remove-lifecycle stale
/triage accepted

[rwnd-bradley]

I have been following the progress of this issue closely since submitting my bug report (scylladb/scylla-manager#3920). Currently, my 5 node scylla cluster experiences a constant throughput of 100-400mbps per node (not sure why it varies so much) due to scylla manager continuously checking the s3 availability (explained in the linked issue), would really like to see this resolved. I was wondering if there is a target release or a roadmap indicating when this issue might be resolved? Thank you for your hard work 🙂

[zimnx]

@rwnd-bradley mentioned issue should be fixed by 1.15.0. Make sure to update Scylla Manager as well to minimal required version.

[rzetelskik]

@rwnd-bradley just to add to @zimnx's comment - the lowest manager version required would be 3.3.3.
The suspected root cause was #1827 and it indeed should've been fixed in 1.15.

If the issue still persists with SO 1.15 and SM >= 3.3.3, feel free to open a new issue in our repository with must-gather attached and we'll investigate.