Bug 1035601 - [RHSC] Remove-brick icon disappears from the UI, when glusterd is killed on the node which was running remove-brick
Summary: [RHSC] Remove-brick icon disappears from the UI, when glusterd is killed on t...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat
Component: rhsc
Version: 2.1
Hardware: Unspecified
OS: Unspecified
high
high
Target Milestone: ---
: RHGS 2.1.2
Assignee: Shubhendu Tripathi
QA Contact: Shruti Sampat
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2013-11-28 07:44 UTC by Shruti Sampat
Modified: 2015-05-13 16:27 UTC (History)
11 users (show)

Fixed In Version: CB11
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2014-02-25 08:06:40 UTC
Target Upstream Version:


Attachments (Terms of Use)
engine logs (14.15 MB, text/x-log)
2013-11-28 08:49 UTC, Shruti Sampat
no flags Details
Attachment for rebalance stop. (209.52 KB, image/png)
2013-12-11 07:07 UTC, RamaKasturi
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Red Hat Product Errata RHEA-2014:0208 0 normal SHIPPED_LIVE Red Hat Storage 2.1 enhancement and bug fix update #2 2014-02-25 12:20:30 UTC
oVirt gerrit 21914 0 None None None Never

Description Shruti Sampat 2013-11-28 07:44:59 UTC
Description of problem:
-------------------------

On a cluster of 4 nodes, with a distribute volume, having bricks on all 4 nodes, remove-brick operation was started for one of the bricks. When remove-brick was in progress, glusterd was killed on the node where the brick being removed, resided. The remove-brick icon in the activities column, disappeared from the UI. The task in the task tab shows UNKNOWN status.

The icon should not disappear from the UI, as long as the task is active on the volume.

Version-Release number of selected component (if applicable):
Red Hat Storage Console Version: 2.1.2-0.25.master.el6_5 
glusterfs 3.4.0.44.1u2rhs

How reproducible:
Tried it once.

Steps to Reproduce:
1. On a distribute volume, with 4 bricks, one on each node of the volume, start removing one brick.
2. Kill glusterd on the same node where the brick was being removed.

Actual results:
The remove-brick icon disappears from the Console.

Expected results:
The icon should not disappear, and it should correctly convey the correct status, could be 'unknown' in this case, because the only participant of the remove-brick operation is unable to return the status ( as glusterd is down on that server ). All other nodes in the cluster would report the task as 'not started'.

The remove-brick status command at the gluster CLI does not display any rows, because non-participating servers are not shown.

Additional info:

Comment 1 Shruti Sampat 2013-11-28 08:49:59 UTC
Created attachment 830098 [details]
engine logs

Comment 4 RamaKasturi 2013-12-11 07:06:16 UTC
When rebalance is stopped from CLI, icon in the activities column gets disappears.

Expected Results : The icon should not disappear, and it should correctly convey the correct status, could be 'unknown' in this case.

This bug needs to be fixed for the above situation as well.

Comment 5 RamaKasturi 2013-12-11 07:07:03 UTC
Created attachment 835136 [details]
Attachment for rebalance stop.

Comment 6 Sahina Bose 2013-12-12 05:19:08 UTC
An ? icon added for Unknown status.

Comment 7 Shruti Sampat 2013-12-17 11:32:39 UTC
The '?' icon is displayed when glusterd is killed on the server on which remove-brick is running, in case of a distributed volume. In case of a distributed-replicate volume, the remove-brick icon is shown, even if glusterd stops running on one of the nodes that has remove-brick running.

The drop-down menu beside the icon shows the status button enabled. Clicking on the Status button shows the status dialog, but the table for remove-brick status is empty.

Either the status button should be disabled, or if it enabled, on clicking on the Status button, a message that status could not be fetched should be shown, instead of the status dialog being shown without any data. Moving this to assigned.

Comment 8 Sahina Bose 2013-12-17 16:18:46 UTC
Shruti, Could you open a separate BZ for comment 7?

Comment 9 Shruti Sampat 2013-12-18 05:52:24 UTC
I had a chat with Shubhendu about this bug, and he was of the opinion that the status button would be enabled, but if you click on it, it would say that "Unable to fetch status details" and that is not what is happening right now. So what is the need for a separate BZ?

Comment 10 Kanagaraj 2013-12-18 08:35:05 UTC
'?' icon is shown when the glusterd goes on a host where the bricks(being removed) are residing for both distribute and distributed-replicate volumes. 


Please open a new bug for showing error popup when the remove-brick status is returned empty. Moving this bug to ON_QA.

Comment 11 Shruti Sampat 2013-12-19 09:37:56 UTC
Opened a new bug for empty status in the dialog - BZ #1044499

Verified this bug by killing glusterd on the participant nodes of remove-brick in case of both distribute and distributed-replicate volumes. Activities icon is replaced by '?' icon and reverts to the remove-brick icon when glusterd is started on those nodes.

Comment 12 RamaKasturi 2013-12-19 09:47:43 UTC
When rebalance is stopped from CLI, a '?' icon comes in the volume activities column and clicking on status column from the drop down next to the '?' icon gives the rebalance status dialog.

Icon from the activities column will be removed after 60 mins or when a new task is started.

Comment 13 Shruti Sampat 2013-12-19 10:00:52 UTC
Marking as verified.

Comment 15 errata-xmlrpc 2014-02-25 08:06:40 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/RHEA-2014-0208.html


Note You need to log in before you can comment on or make changes to this bug.