Bug 1104867

Summary: Paramiko SSH error on multiple cinder volume instances
Product: Red Hat OpenStack Reporter: Rajini Karthik <rajini.karthik>
Component: python-paramikoAssignee: Steve Linabery <slinaber>
Status: CLOSED ERRATA QA Contact: John Walsh <John_walsh>
Severity: high Docs Contact:
Priority: high    
Version: 4.0CC: apevec, dnavale, John_walsh, lhh, markmc, mburns, morazi, rajini.karthik, rhos-maint, sclewis, slinaber, yeylon
Target Milestone: rcKeywords: OtherQA, TestOnly
Target Release: 5.0 (RHEL 7)   
Hardware: x86_64   
OS: Linux   
Whiteboard: MVP
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
: 1136404 (view as bug list) Environment:
Last Closed: 2014-09-02 17:38:31 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1093544, 1108193, 1136404    
Attachments:
Description Flags
paramiko 1.7.6 file with fix none

Description Rajini Karthik 2014-06-04 20:56:21 UTC
Created attachment 902342 [details]
paramiko 1.7.6 file with fix

Description of problem:

Dell Equallogic cinder driver uses paramiko package for transport. When multiple instances of the storage backend is configured, paramiko fails to run the volume manager. 

This issue occurs with paramiko 1.7.6
Error: PID check failed. RNG must be re-initialized after fork()

Version-Release number of selected component (if applicable):
RH-OSP A4

How reproducible:
Configure RH-OSP to use mutiple instances of equallogic arrays as cinder backends


Steps to Reproduce:
1. Use mutiple instances of equallogic or other storage backends that use SSH as cinder backends. Make sure you enable backends and backends are configured in /etc/cinder/cinder.conf
2. Start openstack-cindder-volume service 
3. Check the /var/log/cinder/volume.log file on the cinder controller node
to see the paramiko errors. cinder service-list will show the volume service as down as well

Actual results:
cinder service-list should show the state of the volume services to be down

Expected results:
cinder service-list should show the state of the volume services to be up

Additional info:
Workaround is to upgrade the paramiko package to v 1.7.7 or above, or replace the transport file with fixes and restart cinder volume service
/usr/lib/python2.6/site-packages/paramiko/transport.py

Comment 2 Perry Myers 2014-06-12 18:24:22 UTC
The version of python-paramiko we ship in RHEL OSP 5 is already:
python-paramiko-1.11.3-1.el7ost  

I think the older version you tested with was from core RHEL Server, which is indeed older.

So, please retest with the later version of the package that we ship in OSP 5.  If the problem goes away we can close this as CURRENTRELEASE.

Comment 3 Lon Hohberger 2014-06-12 18:32:03 UTC
The 1.11.3 version is also in the RHEL OSP Beta for RHEL 6 channel.

Comment 10 errata-xmlrpc 2014-09-02 17:38:31 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHBA-2014-1116.html

Comment 11 Rajini Karthik 2014-09-04 16:03:32 UTC
Verified with OSP5 installation that a latest paramiko package is installed. 

On the controller node
[root@mac005056a63bdd ~]# rpm -qa | grep paramiko
python-paramiko-1.11.3-1.el7ost.noarch