Is their any way I can access the host oVirt-engine database and change field of connection to host from IP to FQDN? I tried putting server in maintenance mode.. and even remove one (telling it to ignore gluster) but when I try to remove one node and then hope to re-add to cluster via FQDN .. (so gluster commands stop failing) it keeps saying it can't do that to a cluster gluster node. On Sat, Sep 26, 2020 at 1:41 PM Jeremey Wise <jeremey.wise@gmail.com> wrote:
Another note of color to this.
I can't repair a brick as in gluster it calls bricks by hostname.. and oVirt-engine now thinks of it by IP.
Error while executing action Start Gluster Volume Reset Brick: Volume reset brick start failed: rc=-1 out=() err=['Pre Validation failed on thorst_penguinpages_local_ brick: 172_16_100_103:/gluster_bricks/vmstore/vmstore does not exist in volume: vmstore\nPre Validation failed on odinst_penguinpages_local_ brick: 172_16_100_103:/gluster_bricks/vmstore/vmstore does not exist in volume: vmstore']
On Sat, Sep 26, 2020 at 1:27 PM Jeremey Wise <jeremey.wise@gmail.com> wrote:
I posted that I had wiped out the oVirt-engine.. running cleanup on all three nodes. Done a re-deployment. Then to add nodes back.. though all have entries for eachother in /etc/hosts and ssh works fine via short and long name.
I added nodes back into cluster.. but had to do it via IP to get past error.
Now, if I go to create a volume via the GUI in gluster I get: Error while executing action Create Gluster Volume: Volume create failed: rc=30800 out=() err=["Host 172_16_100_102 is not in 'Peer in Cluster' state"]
Which seems to be related to using IP vs DNS to add gluster volumes https://bugzilla.redhat.com/show_bug.cgi?id=1055928
Question: how do i fix the hosts in cluster being defined by IP vs desired hostname?
-- jeremey.wise@gmail.com
-- jeremey.wise@gmail.com
-- jeremey.wise@gmail.com