Uncontrolled swact on standard system is slow
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
StarlingX |
Won't Fix
|
Low
|
Frank Miller |
Bug Description
Brief Description
-----------------
It takes about 1.5 minutes for 'openstack server list' to recover after a uncontrolled swact on a standard system
Severity
--------
Major
Steps to Reproduce
------------------
Test pre-conditions:
- ensure stx-openstack application is applied via system application-list
- helm list –a shows all releases (eg: garbd, cinder, nova, etc) are in status = DEPLOYED state
- kubectl get pods -n openstack -o wide -w |grep -iEv "running|
- openstack server list --all-projects command works
Test steps:
# On active controller (it was controller-0 for the test):
$ sudo su
$ date;reboot -f
Fri Jul 26 15:38:56 UTC 2019
# On controller-1:
$ while true; do date; timeout 10s openstack server list --all-projects; done
Fri Jul 26 15:39:33 UTC 2019
Fri Jul 26 15:39:43 UTC 2019
Fri Jul 26 15:39:53 UTC 2019
Fri Jul 26 15:40:03 UTC 2019
Fri Jul 26 15:40:13 UTC 2019
Fri Jul 26 15:40:23 UTC 2019
+------
| ID | Name | Status | Networks | Image | Flavor |
+------
| d1aa58d2-
+------
Fri Jul 26 15:40:25 UTC 2019
TC-name:
KPI - uncontrolled swact
Expected Behavior
------------------
Less than 1 minute
Actual Behavior
----------------
Total time for 'openstack server list' to recover after uncontrolled swact: ~1m29s
Reproducibility
---------------
Reproducible
System Configuration
-------
Standard system
Lab-name:
ironpass-33-36
Branch/Pull Time/Commit
-------
stx master as of 20190725T013000Z
Last Pass
---------
STX R1
Timestamp/Logs
--------------
See Test Steps for timestamps
Test Activity
-------------
Regression Testing
tags: | added: stx.regression |
tags: | added: stx.retestneeded |
Changed in starlingx: | |
assignee: | Tao Liu (tliu88) → Daniel Badea (daniel.badea) |
tags: |
added: stx.4.0 removed: stx.3.0 |
Marking as stx.2.0 for now / medium priority given the swact does complete although it's slow.