Fedora Account System
Red Hat Associate
Red Hat Customer
Description of problem (please be detailed as possible and provide log snippests): Currently, OCS in bare metal does not allow external enterprise storage to be used for OSDs. Existing storage resources cannot be used and internal media is supported only. The issue manifests especially in established infra with blade systems and existing iSCSI or FC storage SAN. Version of all relevant components (if applicable): Does this issue impact your ability to continue to work with the product (please explain in detail what is the user impact)? Unable to use existing storage options blocks from leveraging enterprise storage in place and requires substantial different OCP nodes to carry local media as well. In blade systems, this is usually a blocker and would require substantial add-on investment for external OCS and might be not feasible for every infrastructure. Is there any workaround available to the best of your knowledge? Local storage is usually not an option for blade systems. Rate from 1 - 5 the complexity of the scenario you performed that caused this bug (1 - very simple, 5 - very complex)? 4 Can this issue reproducible? See https://bugzilla.redhat.com/show_bug.cgi?id=1823409 Can this issue reproduce from the UI? If this is a regression, please provide more details to justify this: Steps to Reproduce: 1. see https://bugzilla.redhat.com/show_bug.cgi?id=1823409 2. 3. Actual results: OCS cannot consume iSCSI nor FC SAN based storage resources. Expected results: OCS should be able to consume SAN based resources, especially for blade systems. Additional info:
Is this requirement different than Bug 1823409 - OSD needs to recognize and support multipath (edit).
The Rook operator supports multi-path already, I think (just like Neha pointed out) a matter of the ocs-operator to recognize that. If that's the case, we need to close this BZ. Thanks.
Based on comment https://bugzilla.redhat.com/show_bug.cgi?id=1823409#c25, Bug 1823409 solution is not meant to provide the functionality for other than IBM ROKS environment. However, there is a need for it for other environments apart from IBM use, as described above.
Understood, however from a pure Rook perspective, the multi-path is present so there is nothing left to do. We shouldn't be tracking RFE in BZ, we use Jira Epics for this, so I'm moving this to ocs-operator since it's a matter of recognizing it. Thanks.
What does the ocs-operator have to "recognize", exactly? If Rook-Ceph supports them, wouldn't they just work? Moving this out to OCS 4.6 in the meantime.
I don't know but Matthias seems to believe there is something more to do... Maybe different storage classes?
Per my understanding it's solved in 4.5 - https://bugzilla.redhat.com/show_bug.cgi?id=1823409 Travis, can you please confirm?
Yes, this is already working in 4.5. The OCS operator doesn't have any changes required, it's purely recognized during OSD creation. It's expected to work for all environments, not just IBM ROKS. *** This bug has been marked as a duplicate of bug 1823409 ***