Hi, I'm running a two node cluster xen-Cluster with CentOS 5.3. I'm using gfs2 for the cluster filesystem and kernel 2.6.18-128.1.6.el5xen #1 SMP for the Dom0 kernel.
Live migration works fine as long as the DomU guest is running CentOS 5.3 or SLC 4.7 with kernel version 2.6.9-55.EL.cern. But if the SLC 4.7 guest is running the current 2.6.9-78.0.17.EL.cern kernel, the migration freezes without an error message.
Any idea why the DomU kernel has an impact on live migration? Is there a possibility to migrate a DomU guest running kernel 2.6.9-78.0.17.EL.cern? Thanks in advance, Harald
Migration from hosta to hostb hosta /var/log/xen/xend.log .... [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:510) hotplugStatusCallback 1. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices usb. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices vbd. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:160) Waiting for 768. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:496) hotplugStatusCallback /local/domain/0/backend/vbd/17/768/hotplug-status. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:496) hotplugStatusCallback /local/domain/0/backend/vbd/17/768/hotplug-status. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:510) hotplugStatusCallback 1. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:160) Waiting for 5632. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:496) hotplugStatusCallback /local/domain/0/backend/vbd/17/5632/hotplug-status. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:496) hotplugStatusCallback /local/domain/0/backend/vbd/17/5632/hotplug-status. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:510) hotplugStatusCallback 1. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices irq. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices vkbd. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices vfb. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices pci. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices ioports. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices tap. [2009-04-29 09:05:14 xend 5219] DEBUG (DevController:154) Waiting for devices vtpm. [2009-04-29 09:08:04 xend 5219] DEBUG (XendCheckpoint:89) [xc_save]: /usr/lib64/xen/bin/xc_save 23 17 0 0 5 [2009-04-29 09:08:14 xend 5219] INFO (XendCheckpoint:351) Saving memory pages: iter 1 95%^M 1: sent 982654, skipped 355, delta 9327ms, dom0 57%, target 1%, sent 3452Mb/s, dirtied 1Mb/s 412 pages [2009-04-29 09:08:14 xend 5219] INFO (XendCheckpoint:351) Saving memory pages: iter 2 0%^M 2: sent 398, skipped 15, delta 20 ms, dom0 100%, target 0%, sent 652Mb/s, dirtied 24Mb/s 15 pages [2009-04-29 09:08:14 xend 5219] INFO (XendCheckpoint:351) Saving memory pages: iter 3 0%^M 3: sent 0, skipped 15, Start last iteration [2009-04-29 09:08:14 xend 5219] DEBUG (XendCheckpoint:322) suspend [2009-04-29 09:08:14 xend 5219] DEBUG (XendCheckpoint:92) In saveInputHandler suspend [2009-04-29 09:08:14 xend 5219] DEBUG (XendCheckpoint:94) Suspending 17 ... [2009-04-29 09:08:14 xend.XendDomainInfo 5219] DEBUG (XendDomainInfo:1036) XendDomainInfo.handleShutdownWatch [2009-04-29 09:08:14 xend.XendDomainInfo 5219] DEBUG (XendDomainInfo:1036) XendDomainInfo.handleShutdownWatch
hostb /var/log/xen/xend.log .... [2009-04-29 09:08:04 xend 5175] INFO (XendCheckpoint:179) restore hvm domain 11, apic=0, pae=0 [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: boot, val: c [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: fda, val: None [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: fdb, val: None [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: soundhw, val: None [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: localtime, val: None [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: serial, val: pty [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: std-vga, val: None [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: isa, val: None [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: vcpus, val: 1 [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: acpi, val: 1 [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: usb, val: 1 [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: usbdevice, val: None [2009-04-29 09:08:04 xend 5175] DEBUG (image:328) args: k, val: de [2009-04-29 09:08:05 xend 5175] DEBUG (image:388) No VNC passwd configured for vfb access [2009-04-29 09:08:05 xend 5175] DEBUG (XendCheckpoint:198) restore:shadow=0x9, _static_max=0x400, _static_min=0x400, [2009-04-29 09:08:05 xend 5175] DEBUG (balloon:143) Balloon: 29835376 KiB free; need 1065984; done. [2009-04-29 09:08:05 xend 5175] DEBUG (XendCheckpoint:215) [xc_restore]: /usr/lib64/xen/bin/xc_restore 23 11 2 3 1 0 0 [2009-04-29 09:08:05 xend 5175] INFO (XendCheckpoint:351) xc_domain_restore start: p2m_size = 100000 [2009-04-29 09:08:05 xend 5175] INFO (XendCheckpoint:351) Reloading memory pages: 0%
On Wed, Apr 29, 2009 at 09:43:34AM +0200, Harald Schwier wrote:
Hi, I'm running a two node cluster xen-Cluster with CentOS 5.3. I'm using gfs2 for the cluster filesystem and kernel 2.6.18-128.1.6.el5xen #1 SMP for the Dom0 kernel.
Live migration works fine as long as the DomU guest is running CentOS 5.3 or SLC 4.7 with kernel version 2.6.9-55.EL.cern. But if the SLC 4.7 guest is running the current 2.6.9-78.0.17.EL.cern kernel, the migration freezes without an error message.
Any idea why the DomU kernel has an impact on live migration? Is there a possibility to migrate a DomU guest running kernel 2.6.9-78.0.17.EL.cern? Thanks in advance, Harald
This is fedora-xen list, not centos-virt .. :)
This _sounds_ like upstream bug/issue, so you might want to get RHEL support contract and then report it to Redhat..
Or you could try rhelv5-list or xen-devel. I can't really help you with the actual issue..
-- Pasi
what is the architecture of Dom-0 and userdomain? both the same (e.g. 32bit)?
On Thu, Apr 30, 2009 at 9:14 AM, Pasi Kärkkäinen pasik@iki.fi wrote:
On Wed, Apr 29, 2009 at 09:43:34AM +0200, Harald Schwier wrote:
Hi, I'm running a two node cluster xen-Cluster with CentOS 5.3. I'm using gfs2 for the cluster filesystem and kernel 2.6.18-128.1.6.el5xen #1 SMP for the Dom0 kernel.
Live migration works fine as long as the DomU guest is running CentOS 5.3 or SLC 4.7 with kernel version 2.6.9-55.EL.cern. But if the SLC 4.7 guest is running the current 2.6.9-78.0.17.EL.cern kernel, the migration freezes without an error message.
Any idea why the DomU kernel has an impact on live migration? Is there a possibility to migrate a DomU guest running kernel 2.6.9-78.0.17.EL.cern? Thanks in advance, Harald
This is fedora-xen list, not centos-virt .. :)
This _sounds_ like upstream bug/issue, so you might want to get RHEL support contract and then report it to Redhat..
Or you could try rhelv5-list or xen-devel. I can't really help you with the actual issue..
-- Pasi
-- Fedora-xen mailing list Fedora-xen@redhat.com https://www.redhat.com/mailman/listinfo/fedora-xen
The architecture is x86_64 for Dom-0 and Dum-U. (Dell 2950)
what is the architecture of Dom-0 and userdomain? both the same (e.g. 32bit)?
On Thu, Apr 30, 2009 at 9:14 AM, Pasi Kärkkäinen pasik@iki.fi wrote:
On Wed, Apr 29, 2009 at 09:43:34AM +0200, Harald Schwier wrote:
Hi, I'm running a two node cluster xen-Cluster with CentOS 5.3. I'm using gfs2 for the cluster filesystem and kernel 2.6.18-128.1.6.el5xen #1 SMP for the Dom0 kernel.
Live migration works fine as long as the DomU guest is running CentOS 5.3 or SLC 4.7 with kernel version 2.6.9-55.EL.cern. But if the SLC 4.7 guest is running the current 2.6.9-78.0.17.EL.cern kernel, the migration freezes without an error message.
Any idea why the DomU kernel has an impact on live migration? Is there a possibility to migrate a DomU guest running kernel 2.6.9-78.0.17.EL.cern? Thanks in advance, Harald
This is fedora-xen list, not centos-virt .. :)
This _sounds_ like upstream bug/issue, so you might want to get RHEL support contract and then report it to Redhat..
Or you could try rhelv5-list or xen-devel. I can't really help you with the actual issue..
-- Pasi
-- Fedora-xen mailing list Fedora-xen@redhat.com https://www.redhat.com/mailman/listinfo/fedora-xen
-
Harald Schwier -
Harald Schwier -
Pasi Kärkkäinen -
Urs Golla