[vCenter] Cinder-volume service is disabled after reboot cinder-vmware node and lost connectivity with vCenter

Bug #1584862 reported by Ilya Bumarskov
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Fuel for OpenStack
Won't Fix
High
Ilya Bumarskov
Mitaka
Won't Fix
High
Ilya Bumarskov
Newton
Won't Fix
High
Ilya Bumarskov

Bug Description

Fuel 9.0 mos-iso#387 with DVS plugin 3.0.0 build #418

Detailed bug description:
 cinder-volume service is disabled after reboot cinder-vmware node during lost connection with vCenter.

Steps to reproduce:

   - Create env with vCenter and DVS plugin
   - Add following nodes:
         * Controller
         * Controller
         * Controller
         * Cinder-vmware
         * Compute-vmware
   - Deploy env
   - Simulate lost connection with vCenter
   - Reboot cinder-vmware node
   - Restore connection with vCenter
   - Run OSTF "vCenter: Create volume and attach it to instance"

Expected results:
 OSTF is passed

Actual result:
 OSTF has failed. Cinder-volume in down state:
root@node-2:~# cinder service-list
+------------------+------------------------------+----------------+---------+-------+----------------------------+-----------------+
| Binary | Host | Zone | Status | State | Updated_at | Disabled Reason |
+------------------+------------------------------+----------------+---------+-------+----------------------------+-----------------+
| cinder-scheduler | node-2.test.domain.local | nova | enabled | up | 2016-05-23T15:51:36.000000 | - |
| cinder-volume | vcenter@VMwareVcVmdk-backend | vcenter-cinder | enabled | down | 2016-05-23T15:34:33.000000 | - |
+------------------+------------------------------+----------------+---------+-------+----------------------------+-----------------+

tags: added: area-pce-vcenter
Changed in fuel:
importance: Undecided → High
assignee: nobody → Partner Centric Engineering (fuel-partner-engineering)
milestone: none → 9.0
Revision history for this message
Alexander Arzhanov (aarzhanov) wrote :

Logs http://pastebin.com/zuNNFY71

Increase the following parameters is not helping:

vmware_api_retry_count
vmware_task_poll_interval

Revision history for this message
Dina Belova (dbelova) wrote :

Setting to confirmed. MOS Cinder team, please pay attention.

tags: added: area-cinder
Revision history for this message
Ivan Kolodyazhny (e0ne) wrote :

Ilya, provide env to reproduce this bug. MOS-Cinder team doesn't have environment with VMWare

Revision history for this message
Ivan Kolodyazhny (e0ne) wrote :

Actually, it's a normal Cinder behaviour for all backends. If backend driver wasn't initialized on cinder-volume startup, Cinder will never try to do it again.

Fix of this issue should be done in scope of blueprint [1] because it's a improvement and/or feature request.

https://blueprints.launchpad.net/cinder/+spec/driver-initialization-after-fail

Changed in fuel:
milestone: 10.0 → 9.1
Changed in fuel:
status: Confirmed → Won't Fix
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.