Proxmox Clustering fails

F

Fatal Error

New Member
Mar 2, 2010
22
0
1
Hi There

I got some problems while rebuilding our Proxmox Cluster at work.
We had to rebuild the whole thing because we upgraded from Proxmox 1.3 to 1.5 without 1.4 and then got an error with some Ticket authenticaion failed and stuff...
Now i rebuilt these two servers, restored the VMs from the old Cluster to the Master and tried to make the cluster again.

i did it exactly like the Video Tutorial (expect IPs and Passwords of course) and got the ticket authentication failure again.

I have no clue on how to proceed from here and rebuild the cluster on 1.5 and hope you can help me.

So long and thanks in advance
Kevin Schrag
 
log out from the web gui and log in again.

if you still got the issue, check if you can SSH from each other without password.
the ssh keys are in /root/.ssh/ , also take a look into /etc/pve/cluster.cfg

... and why did you choose this user name?
 
tom said:
log out from the web gui and log in again.

if you still got the issue, check if you can SSH from each other without password.
the ssh keys are in /root/.ssh/ , also take a look into /etc/pve/cluster.cfg

... and why did you choose this user name?
Click to expand...
i already checked ssh, was my second action.
cluster.cfg is the same on both servers.
did it fail because there were some VMs running on the master?
 
I assume the solutions is loggin out from the web gui and log in again - can you try with another browser?
 
i simply reinstalled proxmox 1.5 and tried to cluster it without any configurations because i don't want to spend much time on an error when there are no critical systems running on it and got the Ticket failure error again. No running or existing machines, no specific configurations, no storage devices, not even my little bash tool collection is on it at the moment. just proxmox 1.5 out of the box, included the http_proxy export in the two files /etc/profile and /etc/bash.bashrc and then pveca -c on the master and pveca -a -h [masterip] on the node.
now i have:

CID----IPADDRESS----ROLE-STATE--------UPTIME---LOAD----MEM---DISK
1 : xxx.xxx.xxx.251 M S 00:11 0.00 4% 1%
2 : xxx.xxx.xxx.252 N ERROR: Ticket authentication failed - invalid ticket 'root::root::1265451312::0a54f41de08f472d73ea0cce08707079d3e43445'
Click to expand...
thats the output of pveca -l from the master

CID----IPADDRESS----ROLE-STATE--------UPTIME---LOAD----MEM---DISK
1 : 10.26.116.251 M ERROR: Ticket authentication failed - invalid ticket 'root::root::1268129795::5ef8500028ba581d277559b82d1aa110adc07b3b'

2 : 10.26.116.252 N S 02:26 0.00 4% 1%
Click to expand...
thats the output of pveca -l on the node.

what's going on? i can understand the fail at the first try, because there were machines running on the master and such, but now that i have the two servers (both of'em ProLiant DL380 G6) installed out of the box i cannot understand the fail...
 
if you can´t figure it out, go for a support ticket and open SSH/HTTPS access for our support team.
 
udo said:
Hi,
thats very strange... the hostnames of the hosts are different?

Udo
Click to expand...
yeah, the hostnames are different i think.
if i connect via putty it shows me the two different hostnames and i defined them in the proxmoxsetup. do i have to manually set them in the bash via the hostname command?
 
sadly i can't go for a support ticket with port openings because we are a part of a WAN and don't have admin access on the firewalls directly. in addition to this, we don't have any budget for support and stuff so i have to figure it out somehow by myself...

i'll install proxmox 1.3 and try to cluster it because our very first proxmox cluster worked perfectly with 1.3
if this one works, i'll reinstall 1.5 and do the whole install and clustering stuff in a copy paste like manner from your video tutorials in the wiki.
i hope this one will work^^

so long


EDIT:// i installed now 1.3 and clustered the whole story flawless... i will now try to upgrade the servers via 1.4 to 1.5
hopefully it works...
 
Last edited:
so i finally upgraded my Proxmox 1.3 to 1.5 via aptitube upgrade and can't access my eth0.
it doesn't even have that interface or any other eth interface. my only way to connect to my servers now is iLO and i have no idea what to do.
does anyone have some ideas on this point?
 
Fatal Error said:
so i finally upgraded my Proxmox 1.3 to 1.5 via aptitube upgrade and can't access my eth0.
it doesn't even have that interface or any other eth interface. my only way to connect to my servers now is iLO and i have no idea what to do.
does anyone have some ideas on this point?
Click to expand...
Hi,
perhaps your networkdevices will be renamed during the boot?
You can delete the eth0, eth1...-Lines in /etc/udev/rules.d/70-persistent-net.rules
After a reboot you get back an eth0. Perhaps it's not the same eth0 like before (if new driver reconise new hardware, like onboard-lan...)

Udo
 
i had the same problem while creating my cluster :
logoff/logon from the GUI worked
 
Fatal Error said:
i did as you said but got no devices back. do i have to do some stuff manually or should it work automated?
Click to expand...
Hi,
what tells dmesg you about the nic?
You shoul see something like:
Code: 
e1000e: Intel(R) PRO/1000 Network Driver - 1.0.2-k2
e1000e: Copyright (c) 1999-2008 Intel Corporation.
e1000e 0000:04:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
e1000e 0000:04:00.0: setting latency timer to 64
...
ATL1E 0000:06:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
ATL1E 0000:06:00.0: setting latency timer to 64
...
e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX/TX
ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
and some other stuff.
What kind of nic do you use? lspci is your friend.

Udo
 
so, i don't see any eth interface stuff in the dmesg but i'll put it in so you can take a look by yourself.
i also include the lspci, there are the interfaces listed (it's a HP ProLiant DL 380 G6 and i use the built-in onboard devices)

so long and thanks for your effort
Kevin
 

Attachments

Hi there

it's time for some updates...
i installed 1.4 to test if it works on 1.4 and i saw that it didn't work.
i did nothing but clustering, say install 1.4 and cluster it without edit the proxysettings or loading some updates, just out of the box.
i tried relogging in the webinterface and nothing changed.

[hostname]:~# pveca -l
CID----IPADDRESS----ROLE-STATE--------UPTIME---LOAD----MEM---DISK
1 : 10.26.116.251 M S 2 days 22:20 0.00 4% 1%
2 : 10.26.116.252 N ERROR: Ticket authentication failed - invalid ticket 'root::root::1266591755::4bb5e6b8f97abe3b54f12b5a8048196fb9707c9d'
Click to expand...
i get this output when i execute pveca -l on my master

[hostname]:~# pveca -l
CID----IPADDRESS----ROLE-STATE--------UPTIME---LOAD----MEM---DISK
1 : 10.26.116.251 M ERROR: Ticket authentication failed - invalid ticket 'root::root::1269269959::434543f09da0b3b8fd947c4a9a6980f00ebcf55c'

2 : 10.26.116.252 N S 2 days 22:22 0.00 4% 1%
Click to expand...
this one is the node...


i clustered them via putty (SSH) as it shows in the video tutorial.
what to do? i'm completely clueless...

so long
Fatal Error
 
I've seen this error several times while building/adding to a Proxmox Cluster and it was always related to the time on the servers being too far apart from one another.

Are these systems NTP synced?
 
awesome!
thank you, they are finally synching now :-D

I'll try this with 1.5 tomorrow and hopefully it'll work.


So for future references, check the Systemtime (the command is date) and see if they have the same time...
 
You must log in or register to reply here.
Top Bottom