Fedora Account System
Red Hat Associate
Red Hat Customer
Description of problem: ------------------------ as part of perf tests we need to mount the volume to run fio test. While we are unmounting the volume post perf tests, however we are not removing entries from /etc/fstab In the first place, we don't actually need to make entries in /etc/fstab for perf tests as it is a one time test. /etc/fstab entry 172.18.178.65:gluster1 /mnt glusterfs _netdev 0 0 Version-Release number of selected component (if applicable): ========================================== [root@g1-1 glusterfs]# rpm -qa|grep gluster glusterfs-ganesha-3.8.4-52.el7rhgs.x86_64 libvirt-daemon-driver-storage-gluster-3.2.0-14.el7_4.5.x86_64 glusterfs-cli-3.8.4-52.el7rhgs.x86_64 glusterfs-api-3.8.4-52.el7rhgs.x86_64 glusterfs-rdma-3.8.4-52.el7rhgs.x86_64 nfs-ganesha-gluster-2.4.4-17.el7rhgs.x86_64 gluster-zeroconf-avahi-0.1.0-2.el7rhgs.noarch glusterfs-3.8.4-52.el7rhgs.x86_64 glusterfs-libs-3.8.4-52.el7rhgs.x86_64 gluster-nagios-addons-0.2.10-2.el7rhgs.x86_64 glusterfs-server-3.8.4-52.el7rhgs.x86_64 python-gluster-zeroconf-0.1.0-2.el7rhgs.noarch gluster-colonizer-1.0.1-2.el7rhgs.noarch glusterfs-geo-replication-3.8.4-52.el7rhgs.x86_64 vdsm-gluster-4.17.33-1.2.el7rhgs.noarch python-gluster-3.8.4-52.el7rhgs.noarch samba-vfs-glusterfs-4.6.3-9.el7rhgs.x86_64 glusterfs-client-xlators-3.8.4-52.el7rhgs.x86_64 gluster-nagios-common-0.2.4-1.el7rhgs.noarch glusterfs-fuse-3.8.4-52.el7rhgs.x86_64 How reproducible: ============ always Steps to Reproduce: 1.run colonizer to setup gluster, choose any flavor and mount 2.choose to run the perf test 3.check out /etc/fstab and mount command Actual results: ====== can be seen that the volume mount for perf tests job is also being entered into /etc/fstab
This seems to be a quirk of the ansible mount module. Selecting a state of 'mounted' in the play both mounts the volume _and_ creates the fstab entry. - name: Mount default Gluster volume mount: path: /mnt src: "{{ play_hosts[0] }}:{{ default_volname }}" fstype: glusterfs opts: _netdev state: mounted In the tear-down plays, the 'unmounted' state is passed to the mount module, but this _only_ unmounts and does not destroy the fstab entry. - name: Unmount Gluster default volume mount: path: /mnt state: unmounted I would argue that the ansible mount module is behaving incorrectly, but we can easily remedy the problem here by adding a play to remove the fstab entries.
Fix added in upstream merge commit https://github.com/gluster/gluster-colonizer/commit/e8cac3f9fb043bbb35ee612eae0818a6ead03157
on_qa validation: While I see we are still making a entry to /etc/fstab, which I feel is unnecessary for a temporary performance test, However, We are now making sure we are deleting this entry at the end of perf tests. Hence moving to verified. See attached screenshot, where /etc/fstab entries were listed down while perf test was in progress and at the end of perf test. testversion:gluster-colonizer-1.1-2.el7rhgs.noarch.rpm on rhl7.4+gluster-3.8.4-52
Created attachment 1419131 [details] on_qa validation screenshot
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHEA-2018:1317