Bug 1435097 - device remove command hangs when the node on which the device resides is not reachable
Summary: device remove command hangs when the node on which the device resides is not ...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: heketi
Version: cns-3.5
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
: CNS 3.5
Assignee: Raghavendra Talur
QA Contact: krishnaram Karthick
URL:
Whiteboard:
Depends On:
Blocks: 1415600
TreeView+ depends on / blocked
 
Reported: 2017-03-23 07:02 UTC by krishnaram Karthick
Modified: 2018-11-20 06:54 UTC (History)
7 users (show)

Fixed In Version: heketi-4.0.0-6.el7rhgs
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2017-04-20 18:38:24 UTC
Embargoed:


Attachments (Terms of Use)
heketi_logs (1.28 MB, text/plain)
2017-03-23 07:02 UTC, krishnaram Karthick
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Red Hat Product Errata RHEA-2017:1111 0 normal SHIPPED_LIVE heketi bug fix and enhancement update 2017-04-20 22:37:02 UTC

Description krishnaram Karthick 2017-03-23 07:02:41 UTC
Created attachment 1265607 [details]
heketi_logs

Description of problem:
When device remove command is run on a device whose node is not reachable, the command gets hung forever.

There should be a timeout for such cases and cli should error out with a meaningful message.

# heketi-cli device remove 0ce427039fd3bc8f03c43899383fc1ca


[negroni] Started POST /devices/0ce427039fd3bc8f03c43899383fc1ca/state
[negroni] Completed 202 Accepted in 487.498µs
[asynchttp] INFO 2017/03/23 05:58:00 Started job 5ef8ee4fead4ec0f7b9d49cd218deec6
[heketi] INFO 2017/03/23 05:58:00 Replacing brick 2027c32314ae4316bd6877a7f2fc20ae on device 0ce427039fd3bc8f03c43899383fc1ca on node 76c04cd33916422802b3d14e6088ef75
[negroni] Started GET /queue/5ef8ee4fead4ec0f7b9d49cd218deec6
[negroni] Completed 200 OK in 101.92µs


Version-Release number of selected component (if applicable):
heketi-client-4.0.0-3.el7rhgs.x86_64

How reproducible:
1/1

Steps to Reproduce:
1. Have a 4 node CNS setup with some devices and volumes configured on these devices. For eg, node {1..4}, device {1..3}
2. suspend one of the VMs hosting the node. eg, suspend node 1
3. run remove device on node 1 device 1 which has volumes configured on it.

Actual results:
heketi-cli device remove command hangs

Expected results:
command should timeout with proper error message

Additional info:

Comment 2 Humble Chirammal 2017-03-23 07:37:39 UTC
Karthick, whats the behaviour when you issue volume create..etc when one node is down?

Comment 3 Raghavendra Talur 2017-03-31 16:25:32 UTC
This is not a regression. Unfortunately, kubeexec in heketi does not have a timeout and this behaviour will be seen in all ssh/gluster commands. Moving to next release.

Comment 8 krishnaram Karthick 2017-04-12 13:03:11 UTC
Running device disable and device remove on a failed node now works. Moving the bug to verified.

p.s., adding a new node to cns might fail when one of the node is down, filed a separate bug for this case --> https://bugzilla.redhat.com/show_bug.cgi?id=1441675

Comment 10 errata-xmlrpc 2017-04-20 18:38:24 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.

https://access.redhat.com/errata/RHEA-2017:1111


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