Great root cause analysis...
First, a triage note - this bug currently has no package assigned (which
is why the automated tracker stalled with "Package not identified"). It
should be targeted to "cinder" (Ubuntu) so the OpenStack team sees it in
their queue.
Regarding client_cmode.py:
Issuing a single "iscsi-interface-get-iter" without looping over the "next-tag"
element is definitely a bug in multi-node clusters with SLM. The NetApp client
already has pagination helpers (specifically "send_iter_request") used across
other resources like LUNs and igroups, and "get_iscsi_target_details()" clearly
needs to adopt that same iterator pattern.
For an production hotfix while the Cinder SRU is prepared:
You can patch /usr/lib/python3/dist-
packages/cinder/volume/drivers/netapp/dataontap/client/client_cmode.py
to wrap the call with client.send_iter_request('iscsi-interface-get-
iter', ...) instead of the direct single invocation, then restart the
service:
sudo systemctl restart cinder-volume
This will pull all LIFs across all 12 nodes regardless of reporting-node
boundaries.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2166370
Title:
[NetApp] get_iscsi_target_details() does not paginate iscsi-interface-
get-iter, causing iSCSI LIFs to be silently dropped and volume attach
failures with SLM
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2166370/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
