Missing check for cluster partition

Bug #1548679 reported by Peter Sabaini
10
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OpenStack RabbitMQ Server Charm
Fix Released
Wishlist
Unassigned
rabbitmq-server (Juju Charms Collection)
Invalid
Wishlist
Unassigned

Bug Description

The rabbitmq-server charm has a number of nrpe checks included, which is nice. However, it's missing a check that would be of great benefit -- checking for a partitioned cluster.

This could probably be simply done by installing a nrpe check that runs rabbitmqctl cluster_status and matches on something like "\{partitions,\[(.*?)\]\}"

James Page (james-page)
Changed in rabbitmq-server (Juju Charms Collection):
status: New → Triaged
importance: Undecided → Wishlist
James Page (james-page)
Changed in charm-rabbitmq-server:
importance: Undecided → Wishlist
status: New → Triaged
Changed in rabbitmq-server (Juju Charms Collection):
status: Triaged → Invalid
Revision history for this message
Edward Hope-Morley (hopem) wrote :

could also use the update-status hook to report on detected partition

Revision history for this message
James Hebden (ec0) wrote :

Check added to charm here -

https://review.openstack.org/#/c/519594/

I like the idea of updating charm status, too, that would be a trivial reuse of the NRPE check. Potentially, used to kick off repair activities automatically.

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to charm-rabbitmq-server (master)

Reviewed: https://review.openstack.org/519594
Committed: https://git.openstack.org/cgit/openstack/charm-rabbitmq-server/commit/?id=c0b936c7392c554bd3b72cc669484ab0b8960b8d
Submitter: Zuul
Branch: master

commit c0b936c7392c554bd3b72cc669484ab0b8960b8d
Author: James Hebden <email address hidden>
Date: Tue Nov 14 10:34:26 2017 +0100

    Add Nagios check for cluster partitions

    This commit introduces the check_rabbitmq_cluster.py Nagios check, which
    uses the HTTP management API to check for the presence of cluster
    partitions. The check will only be installed if the management_plugin
    charm configuration is set to True, enabling the HTTP API plugin.

    The previous request to use administrator privileges to monitor cluster
    health is no longer required. The create_user and user_exists logic has
    been reworked to accomodate different tags, and the monitoring tag,
    which is respected by the HTTP management API used to monitor cluster
    health, has been used instead of the administrator privilege.

    Also cleans up usage of os.getenv('CHARM_DIR') in several places in the
    file hooks/rabbitmq_server_relations.py

    Change-Id: Ib7eb1afe258931cc917c151a2b6d72dc56d30c95
    Closes-Bug: #1548679

Changed in charm-rabbitmq-server:
status: Triaged → Fix Committed
James Page (james-page)
Changed in charm-rabbitmq-server:
milestone: none → 18.05
David Ames (thedac)
Changed in charm-rabbitmq-server:
status: Fix Committed → Fix Released
tags: added: 4010
Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Change abandoned on charm-rabbitmq-server (master)

Change abandoned by Frode Nordahl (<email address hidden>) on branch: master
Review: https://review.openstack.org/538319
Reason: Marking this review as abandoned due to lack of activity in the past 6 months.
Feel free to restore it again if you want to pick up and continue the work.

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.