Red Hat Bugzilla – Bug 1261290
Snapshot: Disabling shared storage when one or more storage node is down in cluster, should not be allowed
Last modified: 2016-09-17 09:00:00 EDT
Description of problem:
When one or more storage node is down in cluster, disabling shared storage deletes shared storage volume, however when storage node comes up, shared storage volume is created again and Cluster.enable-shared-storage is disabled.
Which will mislead the user.
Version-Release number of selected component (if applicable):
[root@darkknight ~]# rpm -qa | grep glusterfs
Steps to Reproduce:
1. Create shared storage e.g gluster volume set all cluster.enable-shared-storage disable
2. bring down one or two storage nodes, which hosts shared storage bricks
3. disable shared storage e.g gluster volume set all cluster.enable-shared-storage disable
4. Bring up the the storage nodes
When storage nodes comes up, Shared storage volume is created again and gluster volume info shows cluster.enable-shared-storage disable.
This is an existing architecture problem with Glusterd. If volume is deleted when nodes are down then this scenario could occur. This problem will be addressed in Gluterd 2.0. Therefore we are not planning to address this in the current release.
Some aspect of this bug is been addressed in the following bug: