Should wait for all devices to appear?
Bug #1990825 reported by
shoppon
This bug report is a duplicate of:
Bug #1947370: Errors on connect_volume due to race conditions.
Edit
Remove
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
os-brick |
New
|
Undecided
|
Unassigned |
Bug Description
If the iscsi storage has multiple path, os-brick raise exception.
In some scenarios, connect_vol will fail to find the disk due to iscsi get session failure, which will result in missing paths and compromised reliability.
Whether an option should be provided to configure the device that must find all paths, otherwise this scan will fail.
To post a comment you must log in.
You didn't say where this was happening: Controller, Compute, or HCI node.
I'm going to assume this is either Controller or HCI node where you have at least 2 of the following services running: cinder-volume, cinder-backup, glance using cinder as backend, nova-compute.
If that's the case I believe this is a race condition issue that should have been fixed on bug #1947370, though the code also requires setting the same lock path location for the different services for the fix to work.
Going to proceed to close the bug as a duplicate of that one.