I'm not an expert but based on my experience I can recommend you to:
1. Check time sync (NTP/chrony)
2. Check your volumes are configured as described here : https://access.redhat.com/documentation/en-us/red_hat_gluster_storage/3.2/html-single/configuring_red_hat_virtualization_with_red_hat_gluster_storage/index#Configuring_Volumes_Using_the_Command_Line_Interface

Best Regards,
Strahil

On Jan 20, 2019 15:04, Magnus Isaksson <magnus@vmar.se> wrote:

Hello


I have quite some trouble getting Gluster to work.


I have 4 nodes running CentOS and oVirt. These 4 nodes is split up in 2 clusters.

I do not run Gluster via oVirt, i run it stand alone to be able to use all 4 nodes into one gluster volume.


I can add all peers successfully, and i can create an volume and start it with sucess, but after that it starts to getting troublesome.


If i run gluster volume status after starting the volume it times out, i have read that ping-time needs to bee more than 0 so i set it on 30, Still the same problem.


From now on, i can not stop a volume nor remove it, i have to stop glusterd and remove it from /var/lib/gluster/vols/* on all nodes to be able to do anything with gluster.


From time to time when i do a gluster peer status it shows "disconnected" and when i run it again directly after it show "connected"


I get a lot of these errors in glusterd.log

[2019-01-20 12:53:46.087848] W [rpc-clnt-ping.c:223:rpc_clnt_ping_cbk] 0-management: socket disconnected
[2019-01-20 12:53:46.087858] W [rpc-clnt-ping.c:246:rpc_clnt_ping_cbk] 0-management: RPC_CLNT_PING notify failed
[2019-01-20 12:53:55.091598] W [rpc-clnt-ping.c:246:rpc_clnt_ping_cbk] 0-management: RPC_CLNT_PING notify failed
[2019-01-20 12:53:56.094846] W [rpc-clnt-ping.c:246:rpc_clnt_ping_cbk] 0-management: RPC_CLNT_PING notify failed
[2019-01-20 12:53:56.097482] W [rpc-clnt-ping.c:246:rpc_clnt_ping_cbk] 0-management: RPC_CLNT_PING notify failed

What am i doing wrong?


//Magnus