[SOLVED] pvecm nodes -> 1 node IP, all others hostname (upgrade v4/v5)

[pvps1]

I started upgrading from version 4.4. to 5.4 by adding a new host with version 5.4
plan was to migrate old nodes -> new node. upgrade, migrate back, next node etc...


I added the new host (debian stretch + pve repos) with
# pvecm add $ip_old_44_node --use_ssh 1

result:
Nodeid Votes Name
5 1 ada
[...]
12 1 daenerys
14 1 192.168.44.88 (local) <--- this ist the new one
[...]

the NEW 5.4 host is the only thats shown with its IP.
dns resolves forward/backward on all hosts.
hostname/hostname -f etc shows corrects names.
/etc/hosts entry is correct.
in the webgui it appears with its name "rampuri"

when i try to migrate (offline in the case, a test guest) i get
no such cluster node 'rampuri' (500)

I can reproduce this with the gui on a 4.4. host and on the 5.4 host too.

a look at corosyn.conf offers:

node {
name: rampuri
nodeid: 14
quorum_votes: 1
ring0_addr: 192.168.44.88
}

on the other (4.4) hosts it has the hostname:
node {
name: maja
nodeid: 13
quorum_votes: 1
ring0_addr: maja
}

,------------------------------------------------

) what did I miss?
) may I edit the ring0_addr: to ring0_addr: rampuri?

I just don't dare to edit corosync.conf by hand in a production-environment....

any help appreciated, thx
Peter

 

[pvps1]

solved with simply restarting services pveproxy and corosync on the v4.4. machine where the guest-image was.
it stills shows the IP instead of the hostname but moving/migration works