Description of problem: When running "gluster peer status" on several storage nodes of the same trusted storage pool, some nodes show up with their hostnames, some with their IP, despite having the same network settings (hosts, resolv.conf, nsswitch,conf). Version-Release number of selected component (if applicable):2.0 How reproducible: reproducible Steps to Reproduce: 1.run gluster peer status on multipe storage servers of the pool 2. 3. Actual results: [root@rhs1-1 ~]# gluster peer status Number of Peers: 2 Hostname: 192.168.122.12 Uuid: a4e7a8c4-0985-4f11-ad1f-ee730bcd7320 State: Peer in Cluster (Connected) Hostname: rhs1-3 Uuid: bc3609b3-f30e-4231-9b5f-9af878c5ca84 State: Peer in Cluster (Connected) [root@rhs1-2 ~]# gluster peer status Number of Peers: 2 Hostname: 192.168.122.11 Uuid: 8499bba0-f1f0-47c4-ab0e-6236156105b8 State: Peer in Cluster (Connected) Hostname: rhs1-3 Uuid: bc3609b3-f30e-4231-9b5f-9af878c5ca84 State: Peer in Cluster (Connected) [root@rhs1-3 ~]# gluster peer status Number of Peers: 2 Hostname: 192.168.122.11 Uuid: 8499bba0-f1f0-47c4-ab0e-6236156105b8 State: Peer in Cluster (Connected) Hostname: 192.168.122.12 Uuid: a4e7a8c4-0985-4f11-ad1f-ee730bcd7320 State: Peer in Cluster (Connected) Expected results: Hostname row should show hostname in every case. Additional info:
How did you 'gluster peer probe' the storage servers? Was it by IP or by hostname? You can have the hostname updated if you peer-probe the storage server again by hostname. Sometimes it is needed to do the peer-probe from both storage servers: (rhss01) # gluster peer probe server rhss02.example.com (rhss02) # gluster peer probe server rhss01.example.com Does this help?
hmm, of course there is a 'server' too much in the 'gluster peer probe' commands :-/
(In reply to comment #2) > How did you 'gluster peer probe' the storage servers? Was it by IP or by > hostname? > I did add them by hostname: # gluster peer probe rhs1-2 > You can have the hostname updated if you peer-probe the storage server again > by hostname. Sometimes it is needed to do the peer-probe from both storage > servers: > > (rhss01) # gluster peer probe server rhss02.example.com > (rhss02) # gluster peer probe server rhss01.example.com > > Does this help? Yes, this does fix it.
Thanks for confirming. This involves lot of changes internally to make it work consistently. Hence, for now, lets handle this issue using the work around suggested by ndevos.
Could you make sure that the workaround is mentioned in the release notes? Thanks, Patric
I'll write this up for the Red Hat Customer Portal at https://access.redhat.com/knowledge/solutions/201933