devstack plugin: unstack.sh does not kill running zaqar servers

Bug #1504854 reported by Akanksha Srivastava
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
zaqar
Fix Released
Undecided
Eva Balycheva

Bug Description

In a devstack deployment, on running unstack.sh does not kill the zaqar servers. This causes subsequent stack.sh runs to fail to restart zaqar.

Revision history for this message
Eva Balycheva (ubershy) wrote :

I can confirm this bug.

I run zaqar in devstack as plugin.
After executing unstack.sh, there's still zaqar uwsgi master process with it's workers running:
/usr/local/bin/uwsgi --ini /etc/zaqar/uwsgi.conf
Which I can clearly see in the process manager.
This process must be stopped, because if not, zaqar will not start on next execution of stack.sh.

To reproduce bug, set "RECLONE=yes" in DevStack's local.conf and execute:
./stack.sh
./unstack.sh
./stack.sh

Log examples of stack.sh fail: (Reproduced steps twice, I don't know why, but got two different results):
        example1:
            stack.sh, end of log: http://paste.openstack.org/show/mp8mnYyBo8InnI8XFGJi/
     example2:
            stack.sh, end of log: http://paste.openstack.org/show/wIjYLt0SYuLUTPUOm4qa/
            zaqar-wsgi screen, full log: http://paste.openstack.org/show/Uf8ZmhRP69RsAZSCLvNh/
            The logs in this example are more informative.

The problem occurs, because of the nature of uwsgi container. It just can't be stopped like that:
https://github.com/openstack/zaqar/blob/master/devstack/plugin.sh#L253

It must be stopped as described in uwsgi documentation:
https://uwsgi-docs.readthedocs.org/en/latest/Management.html

Thanks for your bug report, Akanksha.

Changed in zaqar:
assignee: nobody → Eva Balycheva (ubershy)
Eva Balycheva (ubershy)
Changed in zaqar:
status: New → Confirmed
Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix proposed to zaqar (master)

Fix proposed to branch: master
Review: https://review.openstack.org/236751

Changed in zaqar:
status: Confirmed → In Progress
Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to zaqar (master)

Reviewed: https://review.openstack.org/236751
Committed: https://git.openstack.org/cgit/openstack/zaqar/commit/?id=39918c82d3637f44be05615efc5c329f4acd16b3
Submitter: Jenkins
Branch: master

commit 39918c82d3637f44be05615efc5c329f4acd16b3
Author: Eva Balycheva <email address hidden>
Date: Mon Oct 19 02:02:04 2015 +0300

    Correctly stop uwsgi process during unstack.sh

    When using Zaqar as plugin in DevStack, after executing unstack.sh,
    there's still Zaqar uwsgi master process with it's workers running:
    /usr/local/bin/uwsgi --ini /etc/zaqar/uwsgi.conf
    This process must be stopped.
    Also running uwsgi process will prevent Zaqar to start on the next
    execution of stack.sh.

    The problem occurs, because of the nature of uwsgi container.
    It just can't be stopped like that:
    https://github.com/openstack/zaqar/blob/master/devstack/plugin.sh#L253

    It must be stopped as described in uwsgi documentation:
    https://uwsgi-docs.readthedocs.org/en/latest/Management.html

    This commit will make uwsgi container process stop correctly on
    unstack.sh execution by using PID of the process as described in uwsgi
    documentation.

    Change-Id: I2f39ea5880c2153fa14a8308d8710969ef6a1a5e
    Closes-Bug: 1504854

Changed in zaqar:
status: In Progress → Fix Committed
Revision history for this message
Thierry Carrez (ttx) wrote : Fix included in openstack/zaqar 2.0.0.0b1

This issue was fixed in the openstack/zaqar 2.0.0.0b1 development milestone.

Thierry Carrez (ttx)
Changed in zaqar:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.