[cinder] not rescanning during VM attach
Hi all, We have a customer problem where LUNs are not able to be scanned in time by cinder, so the operation fails. We fixed the problem in the "create image from volume workflow" by configuring "num_volume_device_scan_tries" in cinder.conf to a larger value. However, this doesn't appear to work for creating a VM and attaching to it, which still always fails. Any configuration or advice here? Thank you, Anthony
On Jul 02, Anthony Galica <anthony.galica@hitachivantara.com> wrote:
Hi all,
We have a customer problem where LUNs are not able to be scanned in time by cinder, so the operation fails.
We fixed the problem in the "create image from volume workflow" by configuring "num_volume_device_scan_tries" in cinder.conf to a larger value.
However, this doesn't appear to work for creating a VM and attaching to it, which still always fails.
Any configuration or advice here?
It would be helpful to have version information, backend storage information, and debug logs. It's unclear if cinder is not scanning sufficiently or if those scans are valid and the host is not yet seeing the LUNs. You can create a bug so we have everything in one place[1] [1] https://bugs.launchpad.net/cinder/+filebug -- Jon
participants (2)
-
Anthony Galica
-
Jon Bernard