Repository navigation
Enhancement: first-class support for parallel/clustered filesystems (IBM Storage Scale, WEKA, Quobyte, ...) as KVM primary storage #14327
Description
Activity
🎯 Triage report
Enhancement request to treat SharedMountPoint on parallel/clustered filesystems (IBM Storage Scale/GPFS, WEKA, Quobyte, GFS2, OCFS2, CephFS) as a first-class, tested configuration, since several assumptions in the current code only hold for NFS or local storage. The author ties this to a concrete bug (#14326) and offers to help test with IBM Storage Scale in a hyperconverged setup.
📊 Assessment
| Dimension | Value | Reasoning |
|---|---|---|
| Type | type:enhancement | Requests documentation, testing coverage, and possibly a new pool capability/flavour rather than fixing a specific defect. |
| Component | component:kvm, component:primary-storage | Concerns KVM agent HA/fencing behavior and SharedMountPoint primary storage. |
| Severity | n/a | Not a bug report. |
| Labels | type:enhancement, component:kvm, component:primary-storage | See above |
| Coding agent | Not suitable | Broad, open-ended ask involving design decisions (new pool flavour/capability, CI scenarios, documentation strategy) and coordination with maintainers — not a well-scoped, self-contained task. |
🔗 Similar issues
- KVM: SharedMountPoint HA heartbeat fails when pool path is a subdirectory of a mounted cluster FS ("Mount point is not a mounted filesystem") #14326 (related) — concrete bug in SharedMountPoint heartbeat logic cited as a motivating example for this enhancement.
- KVM: make storage heartbeat fence action configurable (graceful-reboot / restart-agent / log-only) #13090 (related) — configurable fence action, referenced as related HA/fencing context.
- Disabled the setting
reboot.host.and.alert.management.on.heartbeat.timeoutby default #10111 (related) — default behavior of rebooting host on heartbeat timeout, referenced as related HA/fencing context.
💡 Notes and suggestions
- Natural first steps if maintainers want to pursue this: (1) add a CI/test scenario with a SharedMountPoint pool as a subdirectory of a cluster mount, (2) document recommended HA/fencing settings for hyperconverged deployments, (3) evaluate whether a new pool capability/flag is warranted vs. just fixing filesystem-detection logic (as in KVM: SharedMountPoint HA heartbeat fails when pool path is a subdirectory of a mounted cluster FS ("Mount point is not a mounted filesystem") #14326/PR kvm: fix SharedMountPoint heartbeat when pool path is a subdirectory of a mount #14331).
- The reporter has offered to help test with IBM Storage Scale — worth looping them in if this is picked up.
- Given the scope (design + documentation + possible new abstraction), this is best handled as a discussion/design task among maintainers rather than a single automated PR.
Generated by Daily Issue Triage · sonnet50 92.1K · ◷
Add this agentic workflows to your repo
To install this agentic workflow, run
gh aw add githubnext/agentics/workflows/daily-issue-triage.md@d7c1dc4b72b00607a67caaffdcc216cb64379cf9
ISSUE TYPE
Enhancement Request
COMPONENT NAME
KVM agent, Primary storage (SharedMountPoint), HA/fencing, Documentation
CLOUDSTACK VERSION
4.22.x / 4.23.x and later
SUMMARY
SharedMountPoint is the natural (and often the only) way to run KVM primary storage on
parallel filesystems such as IBM Storage Scale (GPFS), WEKA, Quobyte, and also
on GFS2/OCFS2/CephFS. These setups are increasingly common, often hyperconverged, but
they seem to be rarely covered by tests. Assumptions that only hold for NFS or local
storage keep slipping in, for example:
parallel-FS clusters a simultaneous fence of several hosts destroys the storage quorum
(see Disabled the setting
reboot.host.and.alert.management.on.heartbeat.timeoutby default #10111, KVM: make storage heartbeat fence action configurable (graceful-reboot / restart-agent / log-only) #13090),filesystems.
Suggestions:
one CI scenario with the pool in a subdirectory of a cluster mount),
heartbeat, fencing and mount checks can be adapted per filesystem,
Concrete example: #14326 (SMP heartbeat requires the pool path to be a mount point).
Happy to help test with IBM Storage Scale in a hyperconverged KVM setup.
Related: #12773 (SMP heartbeat), #13090 (configurable fence action), #10111 (reboot on heartbeat timeout disabled by default).