Bug 1475341 - [heketi]: device and node remove should not use 'kill $pid' to kill a brick process [NEEDINFO]
[heketi]: device and node remove should not use 'kill $pid' to kill a brick p...
Product: Red Hat Gluster Storage
Classification: Red Hat
Component: heketi (Show other bugs)
Unspecified Unspecified
unspecified Severity high
: ---
: ---
Assigned To: Raghavendra Talur
krishnaram Karthick
Depends On:
Blocks: 1445448
  Show dependency treegraph
Reported: 2017-07-26 09:13 EDT by krishnaram Karthick
Modified: 2017-08-03 10:13 EDT (History)
6 users (show)

See Also:
Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2017-08-03 10:13:27 EDT
Type: Bug
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---
hchiramm: needinfo? (rtalur)

Attachments (Terms of Use)

  None (edit)
Description krishnaram Karthick 2017-07-26 09:13:55 EDT
Description of problem:
Device remove and node remove operations internally kills the brick process in-order to proceed with the device removal. However, with the brick multiplexing feature in place, killing pid of brick would bring down all the bricks under the pid (ie., all brick processes in the node). 

We need a better way to kill the brick process and proceed with the device or node removal.

Also, any other occurances of pid kill should also be removed (I'm not sure if we do pid kill anywhere else in heketi)

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

How reproducible:
Comment 4 Mohamed Ashiq 2017-08-03 10:13:27 EDT
We are not doing a 'kill' in the heketi for remove device. We were planning to do kill for replace brick but as gluster does that for us, I removed this path. There is no kill in heketi path. Hence closing the bug, if you face the issue related to this please open another bug.

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