when console vm 113 i get vm 115

[offerlam]

Hi all,

This is screwed up!!

I just reinstalled VM 113 with a new centos 7 install. when its done i rebooted like normal. Than i went to proxmox and went for console for vm 113 and now i get vm 115? i can see that because of the hostname. The hostname is the same as for vm 115 and should have been localhost which i centos default? How can this happens?

also

I have 4 VMs now(112 113 116 109) turned off (one of which is also part of the top problem) all having the same IP. Some is a clone of another and some are new mashines. Im working on a problem thats why i have that many with the same ip. they are all turned off...

at least qm status shows them as turned off. but when i ping the ip i get a response?
I think its vm 116 thats is not really turned off.

This is a 3 node cluster running version 3.4-10

I have never seen this before.. i feel the cluster is out of sync somehow. but im no expert..

 

[t.lamprecht]

Hi,

We can not and do not touch the hostname of a VM, so this is weird but should be caused by simply entering it during installation, a small moment of inattention happens to me also when installing a VM alongside some other work I do.

You can change the hostname back and then it should stay like desired.

Is there a process running for those VMs?
# qm list
shows the state, and if you think it may be erroneous use:

ps aux | grep "kvm.*VMID"[code]

replace only VMID with your IDs and look if there is an process running for those (should be a very long process command).

 

[offerlam]

Hi,

We can not and do not touch the hostname of a VM, so this is weird but should be caused by simply entering it during installation, a small moment of inattention happens to me also when installing a VM alongside some other work I do.

You can change the hostname back and then it should stay like desired.

Is there a process running for those VMs?
# qm list
shows the state, and if you think it may be erroneous use:

ps aux | grep "kvm.*VMID"[code]

replace only VMID with your IDs and look if there is an process running for those (should be a very long process command).[/QUOTE]

Hi t.lamprecht 

thanks for answering!

When i see stuff like this the first I question is myself. I too know what you are talking about when you are working with multiple task and you mix em up. I just don't think this is one of them :( when i try and console vm 115 which is the real centreon VM i don't get a picture. It just hangs there. If i do a console on VM 113 i get cli where hostname is centreon IE VM 115. Centos runs with localhost as default and i would need to do a couple of clicks to change it if i wanted. So im pretty sure i didn't change it during install of centos. Bear in mind its not like with ubuntu where it directly asks you for a hostname. You would need to click a GUI button in order to even change it. actually two if im not mistaken.

The VMs in question are spead over two nodes. Node 1 which is called proxmox00 and node 2 which is called proxmox01.

here is the qm list for proxmox00 who is hosting VM 112 113 115 116 where VM 115 is the actual centreon server and 113 is the centos that shows centreon when i console it. 112 116 are the VMs with the same ip where one responds even though all are shutdown

[CODE]
root@proxmox00:~# qm list | grep 112
       112 JQbygOwncloud02      stopped    1024             200.00 0
root@proxmox00:~# qm list | grep 113
       113 200GBOwncloudTemplate02 stopped    2048             200.00 0
root@proxmox00:~# qm list | grep 115
       115 Observium01          running    2048              15.00 686166
root@proxmox00:~# qm list | grep 116
       116 JQByg01Owncloud      stopped    1024             200.00 0
root@proxmox00:~# qm list | grep 113
       113 200GBOwncloudTemplate02 stopped    2048             200.00 0
root@proxmox00:~#

and this is the qm list for proxmox01 which host VM 109 - VM 109 is part of the mashines that has the same ip and responds even though they are all turned off

root@proxmox01:~# qm list | grep 109
       109 JQbygOwncloud03      stopped    2048             200.00 0
root@proxmox01:~#

here is the PS aux output from proxmox00 from the VMs in question:

root@proxmox00:~# ps -aux | grep "kvm.115"
warning: bad ps syntax, perhaps a bogus '-'?
See http://gitorious.org/procps/procps/blobs/master/Documentation/FAQ
root      971795  0.0  0.0   7788   980 pts/0    S+   10:31   0:00 grep kvm.115
root@proxmox00:~# ps -aux | grep "kvm.112"
warning: bad ps syntax, perhaps a bogus '-'?
See http://gitorious.org/procps/procps/blobs/master/Documentation/FAQ
root      971911  0.0  0.0   7788   980 pts/0    S+   10:31   0:00 grep kvm.112
root@proxmox00:~# ps -aux | grep "kvm.113"
warning: bad ps syntax, perhaps a bogus '-'?
See http://gitorious.org/procps/procps/blobs/master/Documentation/FAQ
root      971938  0.0  0.0   7788   980 pts/0    S+   10:31   0:00 grep kvm.113
root@proxmox00:~# ps -aux | grep "kvm.116"
warning: bad ps syntax, perhaps a bogus '-'?
See http://gitorious.org/procps/procps/blobs/master/Documentation/FAQ
root      971942  0.0  0.0   7788   984 pts/0    S+   10:31   0:00 grep kvm.116

and ps output from proxmox01 hosting the last one vm109

root@proxmox01:~# ps aux | grep "kvm.109"
root      863047  0.0  0.0   7788   984 pts/0    S+   10:33   0:00 grep kvm.109
root@proxmox01:~#

do i still make sense or was that too much information? it looks like allot

 

[offerlam]

NEVERMIND!!! IM STUPID!...

sorry for wasting your time!

the reason why its called centreon is because the centos install new ip has a host A record in my DNS.. so centos adopts this hostname which was centreon ... SIGH.. this is getting complex..

 

[t.lamprecht]

Ok good to know that hostname has a reason. Automatic hostnames would be far from nice

Is the still running and so ping'able VM problem also solved now?

PS:
Your ps command was a bit malformed, pleas use

ps aux | grep "kvm.*115"

(note the mssing hyphen on 'aux' and the * after the dot.
If those show a long command like:

qm showcmd 115

would the machine is really still running.

 

[offerlam]

Ok good to know that hostname has a reason. Automatic hostnames would be far from nice

Is the still running and so ping'able VM problem also solved now?

PS:
Your ps command was a bit malformed, pleas use

ps aux | grep "kvm.*115"

(note the mssing hyphen on 'aux' and the * after the dot.
If those show a long command like:
[code
qm showcmd 115
[/code]
would the machine is really still running.

Lets tabel the pingable VM problem for now. I will make a new thread about it if i need help. as you may sense from my posts and problems im vary stressed atm and pressed for time

 

[t.lamprecht]

Yes, a new thread with only the information and problems of the pingable VM and a respective title could be helpful.