drbd problem

[suprstar]

I'm brand new to proxmox, I just inhereted this proxmox 1.7 system which has been running smoothly until now.

I have 2 proxmox nodes running with a dual-primary drbd filesystem. I can migrate some vm's back and forth online, but one I can't. It's a new vm I just created last week, I get an error about the drbdvg-vm--115--disk--1 not existing on node 2. Running 'ls' in /dev/mapper confirms this to be true. How can I get the 2 drbd LVM's back in sync? I am not opposed to completely blowing away the lvm on node2 if it can be recovered from node1, but if there's a faster easier way I'm all ears.

Right now I have all my vm's running on node 1, so I can do pretty much anything to node 2 to get this resolved.

TIA,

 

[udo]

I'm brand new to proxmox, I just inhereted this proxmox 1.7 system which has been running smoothly until now.

I have 2 proxmox nodes running with a dual-primary drbd filesystem. I can migrate some vm's back and forth online, but one I can't. It's a new vm I just created last week, I get an error about the drbdvg-vm--115--disk--1 not existing on node 2. Running 'ls' in /dev/mapper confirms this to be true. How can I get the 2 drbd LVM's back in sync? I am not opposed to completely blowing away the lvm on node2 if it can be recovered from node1, but if there's a faster easier way I'm all ears.

Right now I have all my vm's running on node 1, so I can do pretty much anything to node 2 to get this resolved.

TIA,

Hi,
see here: http://pve.proxmox.com/wiki/DRBD#Recovery_from_communication_failure

Udo

 

[suprstar]

Thank you for the response. I did see that before I came to the forum, but on running

drbdadm secondary r0

on the 'broken' node, I got an error about something holding the resource r0 open, and the command failed and returned some non-zero value. /etc/init.d/drbd restart would give similar results. After a server reboot, drbd did not even start. So I started the service, then the instructions on that page worked. I can migrate vm's back and forth again, my monitoring system stopped reporting errors, and everything is perfectly healthy again. I did reboot the server previously, but I didn't have to explicity start drbd, and any attempts to stop/restart it gave me the same error. I don't know what the difference was this time.

 

[udo]

Thank you for the response. I did see that before I came to the forum, but on running

drbdadm secondary r0

on the 'broken' node, I got an error about something holding the resource r0 open, and the command failed and returned some non-zero value. /etc/init.d/drbd restart would give similar results. After a server reboot, drbd did not even start. So I started the service, then the instructions on that page worked. I can migrate vm's back and forth again, my monitoring system stopped reporting errors, and everything is perfectly healthy again. I did reboot the server previously, but I didn't have to explicity start drbd, and any attempts to stop/restart it gave me the same error. I don't know what the difference was this time.

Hi,
the drbd-resources was held open by lvm.
Use "vgchange -a n vgname" to deactivate the vg (if you have an running vm, or an stale lv you can perhaps not deactivate the vg and must look further).

After that, you can put the drbd-resources in secondary-mode.

To avoid problems i have modified my drbd-startup-script. During stop I deactivating the VGs and change all drbd-resources to secondary. Works for me more stable for reboots.

Udo