Bug 1118319

Summary: failed to attach volumes to instances after configuration change & services restart
Product: Red Hat OpenStack Reporter: Yogev Rabl <yrabl>
Component: openstack-novaAssignee: Nikola Dipanov <ndipanov>
Status: CLOSED WORKSFORME QA Contact: Ami Jeain <ajeain>
Severity: high Docs Contact:
Priority: high    
Version: 5.0 (RHEL 7)CC: eglynn, ndipanov, yeylon, yrabl
Target Milestone: rc   
Target Release: 5.0 (RHEL 7)   
Hardware: x86_64   
OS: Linux   
Whiteboard: storage
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2014-07-22 09:34:45 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:
Attachments:
Description Flags
nova-compute logs none

Description Yogev Rabl 2014-07-10 12:44:00 UTC
Created attachment 917054 [details]
nova-compute logs

Description of problem:
The attachment of volumes failed with the errors that are available in the log file attached. Prior to the error I was running 8 active instances, made a configuration change - increased the number of workers in the Cinder, Nova & Glance services, then restarted the services.   

Ran the command:
# nova volume-attach 6aac6fb6-ef22-48b0-b6ac-99bc94787422 57edbc5c-8a1f-49f2-b8bf-280ab857222d auto

+----------+--------------------------------------+
| Property | Value                                |
+----------+--------------------------------------+
| device   | /dev/vdc                             |
| id       | 57edbc5c-8a1f-49f2-b8bf-280ab857222d |
| serverId | 6aac6fb6-ef22-48b0-b6ac-99bc94787422 |
| volumeId | 57edbc5c-8a1f-49f2-b8bf-280ab857222d |
+----------+--------------------------------------+

cinder list output:
 +--------------------------------------+-----------+---------------+------+-------------+----------+-------------+
|                  ID                  |   Status  |  Display Name | Size | Volume Type | Bootable | Attached to |
+--------------------------------------+-----------+---------------+------+-------------+----------+-------------+
| 57edbc5c-8a1f-49f2-b8bf-280ab857222d | available |   dust-bowl   | 100  |     None    |  false   |             |
| 731a118d-7bd6-4538-a3b2-60543179281e | available | bowl-the-dust | 100  |     None    |  false   |             |
+--------------------------------------+-----------+---------------+------+-------------+----------+-------------+


Version-Release number of selected component (if applicable):
python-cinder-2014.1-7.el7ost.noarch
openstack-nova-network-2014.1-7.el7ost.noarch
python-novaclient-2.17.0-2.el7ost.noarch
openstack-cinder-2014.1-7.el7ost.noarch
openstack-nova-common-2014.1-7.el7ost.noarch
python-cinderclient-1.0.9-1.el7ost.noarch
openstack-nova-compute-2014.1-7.el7ost.noarch
openstack-nova-conductor-2014.1-7.el7ost.noarch
openstack-nova-scheduler-2014.1-7.el7ost.noarch
openstack-nova-api-2014.1-7.el7ost.noarch
openstack-nova-cert-2014.1-7.el7ost.noarch
openstack-nova-novncproxy-2014.1-7.el7ost.noarch
python-nova-2014.1-7.el7ost.noarch
openstack-nova-console-2014.1-7.el7ost.noarch

How reproducible:
100%

Steps to Reproduce:
1. Launch instances
2. Increase the number of workers for the Cinder, Nova & Glance
3. Create a volume
4. Attach the volume to the instance.

Actual results:
The attachment process fail.

Expected results:
The volume should be attached to the instance.

Additional info:

Comment 2 Nikola Dipanov 2014-07-14 16:59:11 UTC
I was not able to reproduce this in any way with the latest puddle.

[root@ndipanov-rhel7-test ~(keystone_demo)]# rpm -q openstack-nova-compute
openstack-nova-compute-2014.1.1-1.el7ost.noarch

I fire up an instance and attach one volume to it. After that I increased the number of API workers in /etc/{nova,cinder}/{nova,cinder}.conf using osapi_compute_workers and osapi_volume_workers options and also in /etc/glance/glance-api.conf using workers option (none of these should even remotely affect anything that is related to attaching volumes thought), and followed it up with 

$ systemctl status openstack-{nova,cinder,glance}-api   

After that I can attach the second volume to the instance without any problems.

What makes me even more suspicious that this is a real bug is that nothing in the attached log suggests any attach failures. There are several detach stack traces in the log though?

Could you please try and reproduce this once again? Otherwise I will close the bug.

Comment 4 Russell Bryant 2014-07-15 13:44:35 UTC
Given Nikola's feedback, I think we need to drop the blocker flag for now.

Comment 5 Yogev Rabl 2014-07-17 06:11:21 UTC
unfortunately, I wasn't able to reproduce it with current version, as well.