Backport of LIO driver related fixes to MOS7.0

Bug #1793205 reported by Andrey Epifanov on 2018-09-18
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Fuel for OpenStack
Roman Lubianyi
tags: added: customer-found
Changed in fuel:
assignee: nobody → Roman Lubianyi (rlubianyi)
Changed in fuel:
milestone: none → 7.0-mu-9
importance: Undecided → High
status: New → Confirmed
summary: - Backport of #1536248 to MOS7.0
+ Backport of LIO driver related fixes to MOS7.0
description: updated

Fix proposed to branch: openstack-ci/fuel-7.0/2015.1.0
Change author: Eric Harney <email address hidden>

Changed in fuel:
status: Confirmed → In Progress

Submitter: Denis V. Meltsaykin <email address hidden>
Branch: openstack-ci/fuel-7.0/2015.1.0

Commit: 38241cb31f8b87950424abb11974ef6434c7beba
Author: Eric Harney <email address hidden>
Date: Mon Sep 24 13:14:08 2018

LIO: Fix terminate_connection AttributeError

"provider_location" is not guaranteed to contain
a value here.

Closes-Bug: #1793205

Change-Id: I964917d13d9415223845ac17eb804ee7faceaf6f
(cherry picked from commit 3e37cb6f8cf489480e264b434290064a61c4a782)

Changed in fuel:
status: In Progress → Fix Committed
Roman Lubianyi (rlubianyi) wrote :

I have tried reproducing the issue on next environment:
MOS7.0 + MU8 updates
Cinder host:
Ubuntu 14.04
cinder-volume 1:2015.1.1-1~u14.04+mos3062
python-cinderclient 1:1.1.2-1~u14.04+mos323
python-rtslib-fb 2.1.45-4~u14.04+mos2
python2.7 2.7.6-8ubuntu0.4

I have performed the next steps:
1) Setup iscsi_helper = lioadm in /etc/cinder/cinder.conf
2) Create Volume and attach it to test VM
3) Reboot the Cinder node without stopping test VM
4) On compute node see in syslog:
root@node-4:~# tail -f /var/log/syslog
<27>Sep 26 09:20:47 node-4 iscsid: connect to failed (Connection refused)
<3>Sep 26 09:22:18 node-4 kernel: [71162.848123] sd 2:0:0:0: rejecting I/O to offline device
<3>Sep 26 11:22:29 node-4 kernel: [78373.809889] sd 2:0:0:0: rejecting I/O to offline device
<3>Sep 26 11:22:29 node-4 kernel: [78373.817055] sd 2:0:0:0: rejecting I/O to offline device

5) On cinder node restore saveconfig.json
root@node-1:~# targetctl restore /etc/rtslib-fb-target/saveconfig.json
Restore failed, 1 errors:
Creating NetworkPortal object failed, skipped

root@node-1:~# netstat -ant |grep 3260
tcp 0 0* LISTEN

root@node-1:~# lsof -i:3260
tgtd 4508 root 7u IPv4 14757 0t0 TCP *:iscsi-target (LISTEN)
tgtd 4508 root 8u IPv6 14758 0t0 TCP *:iscsi-target (LISTEN)

6) Stop and disable tgt service
7) On cinder node repeat restore saveconfig.json
root@node-1:~# targetctl restore /etc/rtslib-fb-target/saveconfig.json

8) Volume successfully reconnected
All next reboots of cinder node with disabled tgt service didn't reproduce the issue. Volume successfully reconnected without losing data in saveconfig.json.

Could you please provide additional information from the customer which will allow me to reproduce the issue on my test environment.

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers