[PVE-User] New node not showing in datacenter

[Emmanuel Kasper]

>>>
>>> Emmanuel
>>>
>> Hi Emmanuel,
>>
>> host102 is an old node, which is not displaying the last one host107 :
>>
>> root at host102:~# pvecm nodes
>>
>> Membership information
>> ----------------------
>>     Nodeid      Votes Name
>>          3          1 10.109.0.101
>>          2          1 10.109.0.102 (local)
>>          1          1 10.109.0.103
>>          4          1 10.109.0.105
>>          5          1 10.109.0.106
>>          6          1 10.109.0.107
>> root at host102:~# pvesh get cluster/config/nodes
>> 200 OK
>> [
>>    {
>>       "name" : "host107",
>>       "node" : "host107",
>>       "nodeid" : "6",
>>       "quorum_votes" : "1",
>>       "ring0_addr" : "10.109.0.107"
>>    },
>>    {
>>       "name" : "host106",
>>       "node" : "host106",
>>       "nodeid" : "5",
>>       "quorum_votes" : "1",
>>       "ring0_addr" : "10.109.0.106"
>>    },
>>    {
>>       "name" : "host102",
>>       "node" : "host102",
>>       "nodeid" : "2",
>>       "quorum_votes" : "1",
>>       "ring0_addr" : "10.109.0.102"
>>    },
>>    {
>>       "name" : "host101",
>>       "node" : "host101",
>>       "nodeid" : "3",
>>       "quorum_votes" : "1",
>>       "ring0_addr" : "10.109.0.101"
>>    },
>>    {
>>       "name" : "host105",
>>       "node" : "host105",
>>       "nodeid" : "4",
>>       "quorum_votes" : "1",
>>       "ring0_addr" : "10.109.0.105"
>>    },
>>    {
>>       "name" : "host103",
>>       "node" : "host103",
>>       "nodeid" : "1",
>>       "quorum_votes" : "1",
>>       "ring0_addr" : "10.109.0.103"
>>    }
>> ]
>>
>>
>>
>> I don't see any error. But not displayed in GUI, and when I try to
>> migrate a VM from host102 to host107 :
>>
>> root at host102:~# qm migrate 126 host107
>> no such cluster node 'host107'
>>
>> As I said in another answer, host105 and host106 were also added a few
>> days ago. I had the same problem with host105 added, but when host106
>> was added, all problem were solved. And now it's host107 turn, but I
>> don't add any other node for now.
>> It seems to be a problem when number of nodes is even and solved when
>> it's odd.

> 
> issue confirmed : I removed node host101, and no more problem with host107
> _______________________________________________


A  bit puzzled here.

Does /etc/pve/.members matches the output of pvecm nodes as you said in
your initial mail ?

The .members file is generated by the pvestatd daemon, so have a look
with journactl -u pvestatd if you see here something suspicious

journalctl -u pve-cluster and journalctl -u corosync might help too.