[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] Reverse migration breaks/freezes VM


  • To: <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Rustedt, Florian" <Florian.Rustedt@xxxxxxxxxxx>
  • Date: Thu, 5 Mar 2009 11:20:16 +0100
  • Delivery-date: Thu, 05 Mar 2009 02:21:21 -0800
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: Acmde/oiNu7UtbA5S4+aL3XyIG0+TA==
  • Thread-topic: Reverse migration breaks/freezes VM

Hello list,

I've got strange phenomenons with migrations. I think it's a bug, but am not 
sure, so i am asking for your help to find this ou, please ;)

First the construct:
Two HP DL180G5(xen0,xen1) with dual-quadcore each, 8G RAM.

On each a dom0 with selfcompiled(because of network drivers) 
gentoo-sources-xen-2.6.27_amd64 and Xen 3.3.0. The installation is completely 
the same (rsynct), only different in ip and hostname.

Both share a drbd0(v8.3.0) device. On it an lvm with one lv per vm.

Now what's the problem:

If i migrate a running vm(mx1-smartnet-de) from xen0 to xen1, it seems to work.
I could do a "xm migrate mx1-smartnet-de xen1" without any interuption while 
having a ssh-connection to the vm and downloading a kernel-source in it.

Then i did a "xm migrate mx1-smartnet-de xen0" on xen1 to move it back again. 
After that, the vm freezes at 100%CPU and that's it! Can't connect to it any 
more, whether via ssh nor via vnc/fb or console. Only "xm destroy" could help, 
"xm shutdown" has no impact.

For debugging, i added the concerning parts of the xend.log from each and the 
kernel-configs, so that you could try to help me find the prob perhaps in the 
logs?

Just after having started the vm: 
xen0_after_start_of_vm.log

Then after the first migration:
xen0_after_migrate_to_xen1_of_vm.log
xen1_after_migrate_to_xen1_of_vm.log

And last, after the reverse:
xen0_after_migrate_back_from_xen1_of_vm.log
xen1_after_migrate_back_from_xen1_of_vm.log

I don't believe it is a drbd/lvm problem, because drbd keeps in "0: 
cs:Connected ro:Primary/Primary ds:UpToDate/UpToDate C r---" and there are no 
lv-concerning log-entries in syslog.

There are of course errors concerning "retry suspend domain failed", but 
there's no hint, WHY this has got failed?

Kind regards, Florian
**********************************************************************************************
IMPORTANT: The contents of this email and any attachments are confidential. 
They are intended for the 
named recipient(s) only.
If you have received this email in error, please notify the system manager or 
the sender immediately and do 
not disclose the contents to anyone or make copies thereof.
*** eSafe scanned this email for viruses, vandals, and malicious content. ***
**********************************************************************************************

Attachment: xen1_after_migrate_to_xen1_of_vm.log
Description: xen1_after_migrate_to_xen1_of_vm.log

Attachment: xen0_after_migrate_back_from_xen1_of_vm.log
Description: xen0_after_migrate_back_from_xen1_of_vm.log

Attachment: xen0_after_migrate_to_xen1_of_vm.log
Description: xen0_after_migrate_to_xen1_of_vm.log

Attachment: xen0_after_start_of_vm.log
Description: xen0_after_start_of_vm.log

Attachment: xen1_after_migrate_back_from_xen1_of_vm.log
Description: xen1_after_migrate_back_from_xen1_of_vm.log

Attachment: config-2.6.27-xen-r5_DOMU
Description: config-2.6.27-xen-r5_DOMU

Attachment: config-2.6.27-xen-r5_DOM0
Description: config-2.6.27-xen-r5_DOM0

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.