2.0 Cluster Migration Fails

[gutter007]

I've installed proxmox 2.0 on two servers, and have configured them into a cluster. I'm trying now to migrate KVM's from server 1 to server 2.

Each time I try and run the migration I get the following error:

May 04 10:48:55 starting migration of VM 201 to node 'server2' (10.10.10.54)
May 04 10:48:55 copying disk images
May 04 10:48:56 ERROR: Failed to sync data - can't migrate local cdrom 'local:iso/CentOS-6.2-x86_64-netinstall.iso'
May 04 10:48:56 aborting phase 1 - cleanup resources
May 04 10:48:56 ERROR: migration aborted (duration 00:00:01): Failed to sync data - can't migrate local cdrom 'local:iso/CentOS-6.2-x86_64-netinstall.iso'
TASK ERROR: migration aborted

The proxmox servers each use local storage. I've tired to find answers on the forums, but I'm still not clear on a few things.

1. Can live or offline migration be done without a shared storage device?
2. If so, what is going wrong with my migration?
3. If not what is the best strategy for maintaining the servers without local storage? Should I delete the cluster, and migrate servers by hand using vzdump and qmrestore?, or are their better options?

Please let me know if you need any more information, and I will send it on.

thanks.
myles.

 

[tom]

the error log shows:

can't migrate local cdrom 'local:iso/CentOS-6.2-x86_64-netinstall.iso'

remove the local:iso and try again.

 

[gutter007]

That fixes that specific error, but then it runs into a similar problem:

May 04 16:35:01 starting migration of VM 101 to node 'bolivar' (10.10.10.54)
May 04 16:35:01 copying disk images
May 04 16:35:01 ERROR: Failed to sync data - can't do online migration - VM uses local disks
May 04 16:35:01 aborting phase 1 - cleanup resources
May 04 16:35:01 ERROR: migration aborted (duration 00:00:00): Failed to sync data - can't do online migration - VM uses local disks
TASK ERROR: migration aborted

Which brings me back to my original points of using clusters with local storage.

 

[tom]

you cannot live migrate if you have local disks. (exception: DRBD)

offline migration works with local disks.

 

[FinnTux]

you cannot live migrate if you have local disks. (exception: DRBD)

How about implementing block migration (-b parameter to migration command)?

 

[tom]

its not implemented it makes not really sense in most uses cases, IMHO. Main reason - takes a very long time if you have big disks.

 

[e100]

its not implemented it makes not really sense in most uses, IMHO. Main reason - takes a very long time if you have big disks.

I wish it was implemented, even submitted an enhancement request to bugzilla here

The live migration with block copy does work, I tested it a few times this week.
Just sucks having to do it manually.

In my opinion, it does not matter if it copies at 50MB/sec or 500MB/sec having this as an option to help avoid downtime is useful to me.
The process could even do a network speed test do some simple math and prompt the user "Based on the speed of your network it will take at least 20 minutes to migrate this VM, are you sure you want to continue?"