[PVE-User] Live-migration slowdown

Julien Groselle julien.groselle at gmail.com
Mon May 27 17:43:56 CEST 2013


Hello everybody,

We have installed 2 servers Gnu/Linux Debian Wheezy.
There are configured in
Cluster<http://pve.proxmox.com/wiki/Proxmox_VE_2.0_Cluster>.
Everything work just fine, but when I want to use live migration we have
this output :

*May 27 14:35:20 starting migration of VM 105 to node 'xxxxx' (IPAddress)*
*May 27 14:35:20 copying disk images*
*May 27 14:35:20 starting VM 105 on remote node 'xxxxx'*
*May 27 14:35:22 starting migration tunnel*
*bind: Cannot assign requested address*
*
*
*May 27 14:35:23 starting online/live migration on port 60000*
*May 27 14:35:23 migrate_set_speed: 8589934592*
*May 27 14:35:23 migrate_set_downtime: 0.1*
*May 27 14:35:25 migration status: active (transferred 39759597, remaining
4794482688), total 8498118656)*
*May 27 14:35:27 migration status: active (transferred 111941833, remaining
1080369152), total 8498118656)*
*May 27 14:35:29 migration status: active (transferred 202341135, remaining
989925376), total 8498118656)*
*May 27 14:35:31 migration status: active (transferred 304080851, remaining
870907904), total 8498118656)*
*May 27 14:35:33 migration status: active (transferred 406818334, remaining
761851904), total 8498118656)*
*May 27 14:35:35 migration status: active (transferred 519978492, remaining
644931584), total 8498118656)*
*May 27 14:35:37 migration status: active (transferred 653085089, remaining
502059008), total 8498118656)*
*May 27 14:35:39 migration status: active (transferred 778259086, remaining
376225792), total 8498118656)*
*May 27 14:35:41 migration status: active (transferred 883049262, remaining
248033280), total 8498118656)*
*May 27 14:35:43 migration status: active (transferred 1026440906,
remaining 92839936), total 8498118656)*
*May 27 14:35:45 migration status: active (transferred 1107951208,
remaining 0), total 8498118656)*
*May 27 14:35:46 migration status: active (transferred 1128351028,
remaining 0), total 8498118656)*
*May 27 14:37:01 migration speed: 82.61 MB/s - downtime 74314 ms*
*May 27 14:37:01 migration status: completed*
*May 27 14:37:04 migration finished successfuly (duration 00:01:44)*
*TASK OK*

In fact the migration work, but it takes more than one minute... I say like
that because I think there are a 60 sec timeout somewhere...
So my problems are :
- After the creation of the migration tunnel I have this error : *bind:
Cannot assign requested address*
Do you have any idea about this issue ?
- At *May 27 14:35:45* the remaining amount of data to sent is 0 but during
one minute, the VM is unreachable !
And i have this in the log file :
> /var/log/syslog :
WARNING: unable to connect to VM 105 socket - timeout after 31 retries
WARNING: unable to connect to VM 105 socket - timeout after 31 retries
WARNING: unable to connect to VM 105 socket - timeout after 31 retries
WARNING: unable to connect to VM 105 socket - timeout after 31 retries
WARNING: unable to connect to VM 105 socket - timeout after 31 retries
WARNING: unable to connect to VM 105 socket - timeout after 31 retries
WARNING: unable to connect to VM 105 socket - timeout after 31 retries
WARNING: unable to connect to VM 105 socket - timeout after 31 retries

During all the time until the "TASK OK"

If one of you have just one idea to resolve this, I will be enjoy to close
the proxmox chapter and start to work on the VM.
Thank you in advance.

Sincerely yours
---
JG
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pve.proxmox.com/pipermail/pve-user/attachments/20130527/f89564bc/attachment-0014.html>


More information about the pve-user mailing list