Problem with cluster cman and quorum

[dghvoip]

I created a cluster and I erroneously changed the name of the machine in /etc/hostname and /etc/hosts on both servers, then I created the cluster and I guess I screwed the machine up, now I tried adding the new cluster member cluster node 1 shows

root@customer:~# pvecm status
cman_tool: Cannot open connection to cman, is it running ?
root@customer:~# /etc/init.d/cman start
Starting cluster:
   Checking if cluster has been disabled at boot... [  OK  ]
   Checking Network Manager... [  OK  ]
   Global setup... [  OK  ]
   Loading kernel modules... [  OK  ]
   Mounting configfs... [  OK  ]
   Starting cman... /usr/sbin/ccs_config_validate: line 186:  4060 Segmentation fault      (core dumped) ccs_config_dump > $tempfile


Unable to get the configuration
corosync [MAIN  ] Corosync Cluster Engine ('1.4.3'): started and ready to provide service.
corosync [MAIN  ] Corosync built-in features: nss
corosync [MAIN  ] Successfully read config from /etc/cluster/cluster.conf
corosync died with signal: 11 Check cluster logs for details
[FAILED]
Stopping cluster:
   Stopping dlm_controld... [  OK  ]
   Stopping fenced... [  OK  ]
   Stopping cman... [  OK  ]
   Unloading kernel modules... [  OK  ]
   Unmounting configfs... [  OK  ]

member 1 shows

root@proxmox-node2:~# pvecm status
Version: 6.2.0
Config Version: 2
Cluster Name: paymcloud
Cluster Id: 55546
Cluster Member: Yes
Cluster Generation: 8
Membership state: Cluster-Member
Nodes: 1
Expected votes: 2
Total votes: 1
Node votes: 1
Quorum: 2 Activity blocked
Active subsystems: 1
Flags:
Ports Bound: 0
Node name: proxmox-node2
Node ID: 2
Multicast addresses: 239.192.xxx.xxx
Node addresses: 217.23.xxx.xxx

I changed the names back to what they were, now another big issue is that machines won't start automatically and when trying to change them from the WebUI I get errors.

I want to know how do I re create the server, if possible or do i have to buy support?, this is a mission critical machine, please advise.

Thanks.

 

[tom]

if you just started probably its best you re-install. but if you know what you are doing you can also fix it.

and yes, we always recommend to go for a support subscription. (see http://forum.proxmox.com/threads/92...ubscriptions-and-why-everybody-should-buy-one)

 

[dghvoip]

if you just started probably its best you re-install. but if you know what you are doing you can also fix it.

and yes, we always recommend to go for a support subscription. (see http://forum.proxmox.com/threads/92...ubscriptions-and-why-everybody-should-buy-one)

1. The steps to fix it would be?... I know what I'm doing just not with this new version.

2. Per CPU socket support, I would ask, what difference does it make for a sever to have 1, 2 or 4 sockets?, kind of a "no no" way to make money no?.

If this offended anyone sorry but that's what I think and I consider unfair, I know this might cost me the answer to my problem, and weird very weird that Proxmox 2 VE which I love has not documented some very important issues, way to drive people to get support as the only solution.

Thanks and Proxmox is great.

 

[dietmar]

If this offended anyone sorry but that's what I think and I consider unfair,

So you think that we should do all those complex tasks for free, and tell everybody that we work for free, because it is considered 'unfair' to work for money? While the software is open source, that does not mean that we are totally brain-dead.

This is the community forum, and you will receive answers if someone wants to help you.

 

[tom]

1. The steps to fix it would be?... I know what I'm doing just not with this new version.

Seems not. The docs says clearly you cannot change the hostname/IP - see "Changing the hostname and IP is not possible after cluster creation." - Therefore I told you to re-install as you are not the expert here.
http://pve.proxmox.com/wiki/Proxmox_VE_2.0_Cluster#Proxmox_VE_Cluster

2. Per CPU socket support, I would ask, what difference does it make for a sever to have 1, 2 or 4 sockets?, kind of a "no no" way to make money no?.

If you are unhappy with our commercial offerings, contact our our sales office.

If this offended anyone sorry but that's what I think and I consider unfair, I know this might cost me the answer to my problem, and weird very weird that Proxmox 2 VE which I love has not documented some very important issues, way to drive people to get support as the only solution.

Thanks and Proxmox is great.

You are wrong, its documented (see above) and you will also find a lot of hints in the forum.

 

[dghvoip]

Seems not. The docs says clearly you cannot change the hostname/IP - see "Changing the hostname and IP is not possible after cluster creation." - Therefore I told you to re-install as you are not the expert here.
http://pve.proxmox.com/wiki/Proxmox_VE_2.0_Cluster#Proxmox_VE_Cluster

Thank you, this was solved re-installing also and I've checked that URL followed in verbatim, but I don´t see anywhere what to do if something goes wrong.

If you are unhappy with our commercial offerings, contact our our sales office.

Ok, Thank you very much again.

You are wrong, its documented (see above) and you will also find a lot of hints in the forum.

Ok, thank you I will do that, alhough I´ve found this post http://forum.proxmox.com/threads/9791-2-1-Cluster-and-node-issues

I try that and no joy either, I see this on my cluster node1

Version: 6.2.0
Config Version: 3
Cluster Name: paycloud
Cluster Id: 27642
Cluster Member: Yes
Cluster Generation: 8
Membership state: Cluster-Member
Nodes: 1
Expected votes: 1
Total votes: 1
Node votes: 1
Quorum: 1
Active subsystems: 5
Flags:
Ports Bound: 0
Node name: 2l23
Node ID: 1
Multicast addresses: 239.xxx.xxx.102
Node addresses: 217.xxx.xxx.82

It seems multicast does work, but I ping the multicast IP with no response, then I login to node2 and -force the addtion but it times out

root@3r19:~# pvecm add 217.xxx.xxx.82 -force
can't create shared ssh key database '/etc/pve/priv/authorized_keys'
The authenticity of host '217.xxx.xxx.82 (217.xxx.xxx.82)' can't be established.
RSA key fingerprint is 8c:20:f0:cd:62:e0:5e:bc:b8:94:cc:ee:75:8e:ce:d4.
Are you sure you want to continue connecting (yes/no)? yes
copy corosync auth key
stopping pve-cluster service
Stopping pve cluster filesystem: pve-cluster.
backup old database
Starting pve cluster filesystem : pve-cluster.
Starting cluster:
   Checking if cluster has been disabled at boot... [  OK  ]
   Checking Network Manager... [  OK  ]
   Global setup... [  OK  ]
   Loading kernel modules... [  OK  ]
   Mounting configfs... [  OK  ]
   Starting cman... [  OK  ]
   Waiting for quorum... Timed-out waiting for cluster
[FAILED]
waiting for quorum...

These servers both have public and private network connections and I´ve tried adding the second node with both and it sits there waiting for quorum.

after a while node1 shows

root@2l23:/var/lib/vz# pvecm nodes
Node  Sts   Inc   Joined               Name
   1   M      8   2012-08-19 18:45:51  2l23
   2   X      0                        3r19

But node2 is still hung waiting for quorum.

So, will I have to re-install that node2 reinstalled?, I rather not, it's a hosted system @ WorldStream so i have no control over that.

I don´t want to turn this into personal attacks or a flame war I simply want some advise where to look for a possible solution.

Thanks and sorry if anyone felt wrong with my post.

 

[tom]

multicast does not work in almost all hosted environments. as your hosting provider.
you write it "seems" working looks not that you are sure.

test multicast - if its not working switch to unicast.

 

[darinschmidt]

2. Per CPU socket support, I would ask, what difference does it make for a sever to have 1, 2 or 4 sockets?, kind of a "no no" way to make money no?.

If this offended anyone sorry but that's what I think and I consider unfair, I know this might cost me the answer to my problem, and weird very weird that Proxmox 2 VE which I love has not documented some very important issues, way to drive people to get support as the only solution.

This is why VMWare has changed their pricing scheme to being based on RAM and not CPUs. 1-4 CPU's can be put into a single server and only one installation is still on that physical server. I dont get it either why it's based on CPU's, not like you need an installation for each. But in a way i understand as then they know you are going to be running many VM's off that puppy, but why should i be charged based on how many VM's you "think" i can fit on a server. I dont even think it should be based on RAM or any physical constraints simply by how many installs you have.

But i'd buy support if it weasnt based on CPUs.... But I'm not posting this to start a pissing match, but dietmar and sometimes tom, from the posts i have read, can be unfriendly. But it isnt their "job" to help us since they need people to buy their $500+ USD subscriptions to keep in business which i cant disagree with. If it wasnt for them, we wouldnt have proxmox. Any project, opensource or not, needs a financial backing so they can live and provide for their families and focus on this software rather than doing it in their spare time.

Best thing to do would be to scour the internet for your answer if it isnt being supplied here, if you dont want to buy the subscription, and share for others to feed on. I'm actually thankful that they even reply to some posts to point people int he right direction and dont ban forums to only paid subscribers. So dont take for granted what they have given to the community.

 

[tom]

two short answers here.

- VMware change backed to CPU socket pricing, vRam is gone.

- yes, "sometimes" our staff help here to support the community ...

Dietmar:
Total Posts 8836
Posts Per Day 3.26

Tom:
Total Posts 7414
Posts Per Day 3.33

And there are countless very helpful community members helping newbies. But I understand that there are always people which do not like what we do here, most times these guys are competitors

 

[darinschmidt]

well son of a gun, good thing there are multi core processors then..... But still though, 1k USD per year for support, isnt all that bad either. You cant even look at oracle support for that kind of price.