OpenStack DBaaS (Trove)

When a backup fails, zombie innobackupex processes are left behind

Bug #1253850 reported by Nikhil Manchanda on 2013-11-22

This bug affects 1 person

Affects		Status	Importance	Assigned to	Milestone
	OpenStack DBaaS (Trove)	Fix Released	Medium	Nikhil Manchanda	OpenStack DBaaS (Trove) 2014.1 "icehouse"

Bug Description

We are calling process.terminate on the pid for the subprocess process, but since shell is True, the only process that gets terminated is the parent shell process. The innobackupex zombie child processes are still alive and left behind, and need to be cleaned up.

Nikhil Manchanda (slicknik) on 2013-11-22

Changed in trove:
assignee:	nobody → Nikhil Manchanda (slicknik)

Revision history for this message

OpenStack Infra (hudson-openstack) wrote on 2013-11-23: Fix proposed to trove (master)

Fix proposed to branch: master
Review: https://review.openstack.org/58066

Changed in trove:
status:	New → In Progress

Nikhil Manchanda (slicknik) on 2013-11-23

Changed in trove:
milestone:	none → icehouse-1

Michael Basnight (hubcap) on 2013-12-03

Changed in trove:
importance:	Undecided → Medium

Thierry Carrez (ttx) on 2013-12-03

Changed in trove:
milestone:	icehouse-1 → icehouse-2

Revision history for this message

OpenStack Infra (hudson-openstack) wrote on 2013-12-06: Fix merged to trove (master)

Reviewed: https://review.openstack.org/58066
Committed: http://github.com/openstack/trove/commit/9e938c63f393879de0309497d7ceef6f9c693b85
Submitter: Jenkins
Branch: master

commit 9e938c63f393879de0309497d7ceef6f9c693b85
Author: Nikhil Manchanda <email address hidden>
Date: Fri Nov 22 17:14:11 2013 -0800

Clean up zombie processes on backup failure

    Ensure that a SIGTERM is sent to the session leader of the process
    group, so that all backup processes are terminated in case of a
    FAILURE. This ensures that no zombie processes are left behind.

Closes bug 1253850

Change-Id: I4bd53c37923f83abe4c51deb95d740656ab21979