New Node Added, now can't login in to web manager

[comcanada]

Hi All,

I added a node to our cluster using "pvecm add x.x.x.x".

The system asked for the root password of the main server, I provided it. After that point, I am no longer able to log in to any of the web gui via https://x.x.x.x:8006

The system seems to recognize the correct root password (I.e. if I enter the wrong password, it tells me the login is incorrect), however, after typing the correct password, the system brings me back to the login page.

Can someone help point me in the right direction?

-Rick

 

[comcanada]

It appears I have made some progress. I shutdown pve-cluster on a server and was able to log in.

Now I'd like to remove/readd it to the cluster but I am unable to get anywhere. Is there an easy way to delete/readd the node in to the cluster?

For example, the "problem" node is pmx8.

root@pmx4:~# pvecm delnode pmx8
node pmx8 does not exist in /etc/pve/cluster.conf
root@pmx4:~# pvecm nodes
Node Sts Inc Joined Name
1 M 468 2012-05-29 19:09:34 pmx1
2 M 468 2012-05-29 19:09:34 pmx11
3 X 488 pmx9
4 M 468 2012-05-29 19:09:34 pmx10
5 M 468 2012-05-29 19:09:34 pmx5
6 M 468 2012-05-29 19:09:34 pmx6
7 M 468 2012-05-29 19:09:34 pmx7
8 X 468 pmx8
9 M 468 2012-05-29 19:09:34 pmx4
10 M 476 2012-05-31 12:28:09 pmx2
root@pmx4:~#

One last update - I noticed that I *CAN* add servers in to the cluster if the hostname & IP address have never been added before. I guess now all I need to figure out is how to delete old hostnames/IPs from the system.

 

[dietmar]

Simply add with --force flag.

 

[comcanada]

Hi,

Thanks for the quick response.

'pvecm nodes' gives me a list of active nodes - that list appears correct.

However, the web gui still shows two servers that were previously deleted... is there some way I can fix that?

-Rick

 

[tom]

remove the corresponding directories in /etc/pve/nodes

but be careful here.

 

[e100]

I had this same issue this week. Removed some nodes but they still showed up in the GUI.
Once I moved/deleted the vm config files from the removed node directory the node dissapeared from the GUI.

 

[dietmar]

Yes, you need to remove remaining data from the file system in /etc/pve/nodes/[NODENAME]/

 

[comcanada]

Excellent, that helped considerably!!

Okay - another one.

I have a pmx server that shows up in the GUI as being red. Summary tab shows correct uptime, cpu's IO relay, ram usage etc. No graphic stats.
I can even start and stop the VEs, though they do not show their DNS names.

If I try to migrate a VM to another server, I get the error "Cluster not ready - no quorum? (500)".

This particular server also does not show in the nodes list when issuing "pvecm nodes", however, there is a directory for it in /etc/pve/nodes.

How can I get this to show up as "green" and bring this back in to the cluster?