[ClusterLabs] Configure a resource to only run a single instance at all times

Andrei Borzenkov arvidjaar at gmail.com
Mon Oct 29 17:19:51 UTC 2018


29.10.2018 20:04, jm2109384 at gmail.com пишет:
> Hi Guys,
> 
> I'm a new user of pacemaker clustering software and I've just configured a
> cluster with a single systemd resource. I have the following cluster and
> resource configurations below. Failover works perfectly between the two
> nodes however, i wanted to have a constraint/rule or a config that will
> ensure that my resource has a single instance running on the cluster at all
> times. I'd like to avoid the situation where the resource gets started
> manually and ends up running on both cluster nodes.  Hoping to get your
> advice on how to achieve this. Thanks in advance.

pacemaker does one time probe on each node when pacemaker is started.
This covers the case when resource was manually started before
pacemaker. You can enable monitor on inactive nodes which should also
detect if resource was started outside of pacemaker. But note that it
leaves you some window (up to monitoring interval) when multiple
instances may be up on different nodes until pacemaker is aware of it.

> 
> ----
> Cluster Name: cluster1
> Corosync Nodes:
> node1 node2
> Pacemaker Nodes:
> node1 node2
> 
> Resources:
> Resource: app1_service (class=systemd type=app1-server)
>   Operations: monitor interval=10s (app1_service-monitor-interval-10s)
>               start interval=0s timeout=120s
> (app1_service-start-interval-0s)
>               stop interval=0s timeout=120s (app1_service-stop-interval-0s)
>               failure interval=0s timeout=120s
> (app1_service-failure-interval-0s)
> 
> Stonith Devices:
> Fencing Levels:
> 
> Location Constraints:
> Ordering Constraints:
> Colocation Constraints:
> Ticket Constraints:
> 
> Alerts:
> No alerts defined
> 
> Resources Defaults:
> resource-stickiness: 100
> migration-threshold: 1
> failure-timeout: 120s
> Operations Defaults:
> No defaults set
> 
> Cluster Properties:
> cluster-infrastructure: corosync
> dc-version: 1.1.18-11.el7_5.3-2b07d5c5a9
> have-watchdog: false
> last-lrm-refresh: 1540829641
> no-quorum-policy: ignore
> stonith-enabled: false
> symmetric-cluster: true
> 
> Quorum:
>   Options:
> 
> 
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> https://lists.clusterlabs.org/mailman/listinfo/users
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
> 



More information about the Users mailing list