[LU-16472] Determine setting of invalid faillocs Created: 13/Jan/23 Updated: 13/Jan/23 |
|
| Status: | Open |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Minor |
| Reporter: | Oleg Drokin | Assignee: | WC Triage |
| Resolution: | Unresolved | Votes: | 0 |
| Labels: | None | ||
| Severity: | 3 |
| Rank (Obsolete): | 9223372036854775807 |
| Description |
|
A somewhat common problem during interop testing (but also due to code mistakes and such) is attempting to set a failloc that does not exist. Right now this is not noticed and then the test fails or falsely succeeded because the expected action does not happen. We need to come up with a way to track this and return error from an attempt of setting the non-existing failloc so that the test scripts could do some sensible thing like failing with very sensible error message or skip the test or some such.
How to implement such tracking is up to a debate - we cannot compile the list of valid failloc values into userspace tools because they could be a mix and match with lustre versions. I guess we can have a separate "Checking" module for this only used during debugging or we can hardcode this into like obdclass or some such, collected from a new macro that would need to be created to track all defined values. |