C
Chris Rivera
Guest
I have had an issue with node 8 and saw that it was offline, but had issues with cman on all other nodes.
I could not stop or restart cman services on any of the nodes.... after not being able to solve the issue with cman, and not being able to login the web gui ( which took long to load... and would not accept the password, WIERD password failed)
i could not ssh into the box but could access via pluggin in KVM. Stilll i could not get cman to stop or start and decided to reboot node 8.
Node 8 gave errors talking about leaks and hex errors related to the network.. once this node rebooted.. All issues with cman, quorum, all cleared up on all nodes. The web gui instantly became accessible accepted the password and is good to go.
How could 1 node cause the whole cluster to not function properly? I would expect that node to drop off the cluster and function as normal but not cause any issues with the other nodes in the cluster.
This is an issue... if i have 20-30 nodes and i have 1 node for some magical reason crap out like node 8 did, i cannot have the cluster in a state where i cannot manage it via GUI.
Would like to know how to avoid another issue like this. thanks in advance
I could not stop or restart cman services on any of the nodes.... after not being able to solve the issue with cman, and not being able to login the web gui ( which took long to load... and would not accept the password, WIERD password failed)
i could not ssh into the box but could access via pluggin in KVM. Stilll i could not get cman to stop or start and decided to reboot node 8.
Node 8 gave errors talking about leaks and hex errors related to the network.. once this node rebooted.. All issues with cman, quorum, all cleared up on all nodes. The web gui instantly became accessible accepted the password and is good to go.
How could 1 node cause the whole cluster to not function properly? I would expect that node to drop off the cluster and function as normal but not cause any issues with the other nodes in the cluster.
This is an issue... if i have 20-30 nodes and i have 1 node for some magical reason crap out like node 8 did, i cannot have the cluster in a state where i cannot manage it via GUI.
Would like to know how to avoid another issue like this. thanks in advance