Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1049388

Summary: Creating an image caused disk trashing and non-responsive browser.
Product: Red Hat OpenStack Reporter: Eduardo Warszawski <ewarszaw>
Component: openstack-cinderAssignee: Flavio Percoco <fpercoco>
Status: CLOSED INSUFFICIENT_DATA QA Contact: Dafna Ron <dron>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 4.0CC: eglynn, eharney, ewarszaw, fpercoco, iheim, yeylon
Target Milestone: ---   
Target Release: 5.0 (RHEL 7)   
Hardware: Unspecified   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2014-02-05 08:28:51 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
cinder volumes log
none
horizon log none

Description Eduardo Warszawski 2014-01-07 14:09:57 UTC
Description of problem:
During glance image creation from dashboard using a qcow2 source, the browser started intensive disk operations caused the client computer to freeze.

Version-Release number of selected component (if applicable):
openstack-glance-2014.1-0.1.b1.el6.noarch

How reproducible:
Upload an qcow2 image in arora browser.

Actual results:
The image is not created.
The local disk is heavily used, the client is unusable.

Expected results:
The image should be created.

Additional info:
openstack-packstack-2013.2.1-0.27.dev936.el6.noarch
openstack-glance-2014.1-0.1.b1.el6.noarch
arora 0.11.0

No traces of the image name in any of the logs.

Comment 1 Flavio Percoco 2014-01-09 12:25:00 UTC
Did you check horizon logs?

Sounds more like a horizon bug to me. Could you please double check?

Comment 4 Eduardo Warszawski 2014-01-15 13:02:10 UTC
Created attachment 850486 [details]
cinder volumes log

Comment 5 Eduardo Warszawski 2014-01-15 13:03:26 UTC
Created attachment 850489 [details]
horizon log

Comment 6 Flavio Percoco 2014-01-16 13:27:11 UTC
Were you trying to create an image or a volume based on an existing image? You attached cinder volume logs but filed the bug under openstack-glance. Could you please clarify?

Cinder logs show an error when trying to create a volume from an image bigger than the available space. Is that error related to what you were trying to do? or did that happen before you hit this issue?

Also, horizon logs show lot of tokens requests. That doesn't sound right.

Based on that, could you please identify which process was I/O intensive?

Please, make sure this kind of information is part of the bug report. Thanks! :)

Comment 7 Eduardo Warszawski 2014-01-21 13:09:21 UTC
(In reply to Flavio Percoco from comment #6)
> Were you trying to create an image or a volume based on an existing image?
A volume based on an existing (qcow2) image.

> You attached cinder volume logs but filed the bug under openstack-glance.
> Could you please clarify?
The image seems to be OK. qemu-img convert is invoked by cinder and continues forever.
The convert is converting the image to raw after a temporal copy was done.
> 
> Cinder logs show an error when trying to create a volume from an image
> bigger than the available space. Is that error related to what you were
> trying to do? or did that happen before you hit this issue?
It has happened before and is not related to the actual issue.
> 
> Also, horizon logs show lot of tokens requests. That doesn't sound right.
> 
You are rigth.

> Based on that, could you please identify which process was I/O intensive?
> 
> Please, make sure this kind of information is part of the bug report.
> Thanks! :)

Comment 8 Flavio Percoco 2014-01-22 09:58:46 UTC
(In reply to Eduardo Warszawski from comment #7)
> (In reply to Flavio Percoco from comment #6)
> > Were you trying to create an image or a volume based on an existing image?
> A volume based on an existing (qcow2) image.
> 
> > You attached cinder volume logs but filed the bug under openstack-glance.
> > Could you please clarify?
> The image seems to be OK. qemu-img convert is invoked by cinder and
> continues forever.
> The convert is converting the image to raw after a temporal copy was done.

This seems like a cinder bug, then. Did you try to convert that image manually and see what happens?

Do you have a running environment I could ssh in?

Moving this to cinder for now.