New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE [UPDATE]

[martin]

Hi all,

[UPDATE]

Just updated some new packages to the pvetest repository (bugfixes and some new features):

• Fixed drive order bug (KVM)
• Allow up to 4094 bridges
• update to the latest vzctl
• Use option –sparse for qmigrate

__
We just uploaded new 2.6.24 and 2.6.32 kernels to our pvetest repository.
(deb http://download.proxmox.com/debian lenny pvetest).

Both kernel branches are using qemu-kvm 0.12.3 and gPXE 1.0. Network boot should work now without issues, tested also with Microsoft WDS from Win2008 r2.

Both kernels are tested in our lab and I encourage everybody to test also.

Change log 2.6.24 (KVM and OpenVZ):

• update to qemu-kvm-0.12.3
• include gPXE 1.0 network boot
• update e1000e to e1000e-1.1.2.tar.gz
• update igb to igb-2.1.9.tar.gz
• update e1000 to e1000-8.0.19.tar.gz
• update bnx2 to netxtreme2-5.2.50.tar.gz
• update tg3 to tg3-3.105h.tar.gz
• kernel update (latest 2.6.24)
• pveversion -v
  

  pve-manager: 1.5-9 (pve-manager/1.5/4728)
  running kernel: 2.6.24-11-pve
  proxmox-ve-2.6.24: 1.5-22
  pve-kernel-2.6.24-11-pve: 2.6.24-22
  qemu-server: 1.1-14
  pve-firmware: 1.0-4
  libpve-storage-perl: 1.0-11
  vncterm: 0.9-2
  vzctl: 3.0.23-1pve9
  vzdump: 1.2-5
  vzprocps: 2.0.11-1dso2
  vzquota: 3.0.11-1
  pve-qemu-kvm: 0.12.3-1

Change log 2.6.32 (KVM only):

• update to qemu-kvm-0.12.3
• include gPXE 1.0 network boot
• kernel update (latest 2.6.32)
• pveversion -v
  

  pve-manager: 1.5-9 (pve-manager/1.5/4728)
  running kernel: 2.6.32-2-pve
  proxmox-ve-2.6.32: 1.5-7
  pve-kernel-2.6.32-2-pve: 2.6.32-7
  qemu-server: 1.1-14
  pve-firmware: 1.0-4
  libpve-storage-perl: 1.0-11
  vncterm: 0.9-2
  vzctl: 3.0.23-1pve9
  vzdump: 1.2-5
  vzprocps: 2.0.11-1dso2
  vzquota: 3.0.11-1
  pve-qemu-kvm: 0.12.3-1
  ksm-control-daemon: 1.0-3

We appreciate all feedback and bug reports in this forum.

Best regards,

Martin

 

[valshare]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

Hello Martin,

i have updted one server to the latest 2.6.24 Kernel. We have a time critial VM running under Windows XP. The Software, that runs on this machine needs a usb hardware dongle for the copy protected application. Till the Kernel 2.6.24-10 the machine runs stable with no problems. Now i have installed the latest 2.6.24-11 Kernel and the latest tools.

Now, the machine has problem to find the usb hardware dongle. To attache the hardware dongle to the xp machine, we use anywhereusb and Belkin Network USB Hub that works great till the latest pvetest.

How can i help to "debug" the problem?!

Regards, Valshare

pve-manager: 1.5-8 (pve-manager/1.5/4674)
running kernel: 2.6.24-11-pve
proxmox-ve-2.6.24: 1.5-22
pve-kernel-2.6.24-11-pve: 2.6.24-22
pve-kernel-2.6.24-7-pve: 2.6.24-11
qemu-server: 1.1-12
pve-firmware: 1.0-4
libpve-storage-perl: 1.0-10
vncterm: 0.9-2
vzctl: 3.0.23-1pve8
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.12.3-1

so, now i have done a downgrade to kernel 2.6.24-10 and all works fine again.

pve-manager: 1.5-8 (pve-manager/1.5/4674)
running kernel: 2.6.24-10-pve
proxmox-ve-2.6.24: 1.5-21
pve-kernel-2.6.24-7-pve: 2.6.24-11
pve-kernel-2.6.24-10-pve: 2.6.24-21
qemu-server: 1.1-12
pve-firmware: 1.0-4
libpve-storage-perl: 1.0-10
vncterm: 0.9-2
vzctl: 3.0.23-1pve8
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.11.1-2

 

[prometheanfire]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

Aparently there is a bug that noone on the kvm dev team is working on for drives over 1TB and virtio. Just an fyi on that. Also, does the new kernel have nfs support built in? I installed nfs-kernel-server but I get failures.

 

[dietmar]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

Also, does the new kernel have nfs support built in? I installed nfs-kernel-server but I get failures.

All our kernels support nfs - its a bug in the init.d script which prevents starting the server on 2.6.32 (but nfs on 2.6.32 is known to have other bugs). So I would simple use nfs-user-server

 

[prometheanfire]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

All our kernels support nfs - its a bug in the init.d script which prevents starting the server on 2.6.32 (but nfs on 2.6.32 is known to have other bugs). So I would simple use nfs-user-server

I actually updated to their parched nfs-kernel-server-6lenny2 and it works now.

 

[euromancer]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

Seems that this upgrade (2.6.32-2 and kvm 0.12.3) help ALOT on my high CPU load on the host when guest is doing alot of IO.
So, keep up the good work

 

[valshare]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

Hi, no one has the same problem with Network to USB Hub? Can anyone test it please?!
Martin, can you verify the problem?

 

[bert64]

How have the KVM 0.12.3 packages been compiled?
I ask because, i manually compiled 0.12.3 before, and when compiled against the version of glibc and kernel headers shipping with proxmox i had serious problems with network stability (under heavy load such as multiple instances of bittorrent the network would go down in the guest, requiring me to ifconfig it up/down), however having recompiled kvm using a glibc compiled against 2.6.32 kernel headers the issues have gone away.

 

[prometheanfire]

If you are using virtio, that is a known bug and you should still see the problem under newer kernels with heavy load, if you are not, lucky you

 

[dietmar]

however having recompiled kvm using a glibc compiled against 2.6.32 kernel headers the issues have gone away.

We use standard debian glibc (lenny)

 

[meto]

During offline migration i have this error:

/usr/sbin/qmigrate 10.10.1.20 104
Apr 30 18:02:50 starting migration of VM 104 to host '10.10.1.20'
Apr 30 18:02:50 copying disk images
rsync: --sparse cannot be used with --inplace
rsync error: syntax or usage error (code 1) at main.c(1441) [client=3.0.3]
Apr 30 18:02:50 found stale volume copy 'local:104/vm-104-disk-1.qcow2' on host '10.10.1.20'
Apr 30 18:02:50 Failed to sync data - command '/usr/bin/rsync --sparse --progress --inplace --no-whole-file /var/lib/vz/images/104/vm-104-disk-1.qcow2 root@10.10.1.20:/var/lib/vz/images/104/vm-104-disk-1.qcow2' failed with exit code 1
Apr 30 18:02:50 migration aborted
VM 104 migration failed -

Kernel 2.6.32 pvetest

Maybe some ideas to fix it: http://www.finalcog.com/rsync-vm-sparse-inplace-kvm-vmware

 

[dietmar]

Maybe some ideas to fix it: http://www.finalcog.com/rsync-vm-sparse-inplace-kvm-vmware

Thanks for that link - will try to fix it that way.

 

[dietmar]

just uploaded a fix to the pvetest repository - please can you test again?

 

[meto]

just uploaded a fix to the pvetest repository - please can you test again?

OK.

/usr/sbin/qmigrate 10.10.1.20 104
May 03 21:16:41 starting migration of VM 104 to host '10.10.1.20'
May 03 21:16:41 copying disk images
vm-104-disk-2.raw

rsync status: 32212254720 100% 93.71MB/s 0:05:27 (xfer#1, to-check=0/1)

sent 983119 bytes received 1966111 bytes 5732.23 bytes/sec
total size is 32212254720 speedup is 10922.26
vm-104-disk-1.qcow2

rsync status: 2778464256 100% 50.58MB/s 0:00:52 (xfer#1, to-check=0/1)

sent 2778803505 bytes received 31 bytes 51940253.01 bytes/sec
total size is 2778464256 speedup is 1.00
May 03 21:26:15 migration finished successfuly (duration 00:09:35)
VM 104 migration done

--edit
Should it improve performance of migrating raw volumes? 30gb raw, empty disk:

/usr/sbin/qmigrate 10.10.1.20 104
May 03 21:16:41 starting migration of VM 104 to host '10.10.1.20'
May 03 21:16:41 copying disk images
vm-104-disk-2.raw

rsync status: 32212254720 100% 93.71MB/s 0:05:27 (xfer#1, to-check=0/1)

sent 983119 bytes received 1966111 bytes 5732.23 bytes/sec
total size is 32212254720 speedup is 10922.26
vm-104-disk-1.qcow2

rsync status: 2778464256 100% 50.58MB/s 0:00:52 (xfer#1, to-check=0/1)

sent 2778803505 bytes received 31 bytes 51940253.01 bytes/sec
total size is 2778464256 speedup is 1.00
May 03 21:26:15 migration finished successfuly (duration 00:09:35)
VM 104 migration done

At first i waited long for rsync status for raw disk, maybe there's a problem. Stale copy maybe should be created via remote command or is it piped through SSH?

 

[valshare]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

looks like my posts are ignored from the developers

Hi, no one has the same problem with Network to USB Hub? Can anyone test it please?!
Martin, can you verify the problem?

 

[tom]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

not ignored, but it looks that there is not quick solution. can you give me (again) complete instructions how to reproduce the issue?

 

[valshare]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

Hi Tom,

thanx for reply.

We use a this and this Network USB Hub for a time critical Application that use a USB dongle for copy protection. Both Hubs runs without any problem.

After upgrade the Proxmox Testserver from

pve-manager: 1.5-8 (pve-manager/1.5/4674)
running kernel: 2.6.24-10-pve
proxmox-ve-2.6.24: 1.5-21
pve-kernel-2.6.24-7-pve: 2.6.24-11
pve-kernel-2.6.24-10-pve: 2.6.24-21
qemu-server: 1.1-12
pve-firmware: 1.0-4
libpve-storage-perl: 1.0-10
vncterm: 0.9-2
vzctl: 3.0.23-1pve8
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.11.1-2

to

proxmox01:~# pveversion -v
pve-manager: 1.5-9 (pve-manager/1.5/4728)
running kernel: 2.6.24-11-pve
proxmox-ve-2.6.24: 1.5-22
pve-kernel-2.6.24-11-pve: 2.6.24-22
pve-kernel-2.6.24-7-pve: 2.6.24-11
qemu-server: 1.1-14
pve-firmware: 1.0-4
libpve-storage-perl: 1.0-13
vncterm: 0.9-2
vzctl: 3.0.23-1pve11
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.12.3-1

The Network to USB Hub don´t work anymore. The Connection to the Network USB Hub lost every one(?) minute and the application exists. After downgrade the testserver to the "old" kernel / proxmox tool all runs fine again.

Think there is a bug in qemu-server (?). How can i help to debug the problem!?

Regards,

Valle

 

[tom]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

we have no such hardware/software/dongle here and therefore we cannot test and reproduce this. so if you want that we try to reproduce the issue here we need all components.

 

[valshare]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

i can send you a bridge for testing, if you want?! Or if you want, i can give you access to the Server / Client.

 

[dietmar]

Re: New Proxmox VE Kernels 2.6.24 and 2.6.32, including KVM 0.12.3 and gPXE network b

The Network to USB Hub don´t work anymore. The Connection to the Network USB Hub lost every one(?) minute and the application exists. After downgrade the testserver to the "old" kernel / proxmox tool all runs fine again.

Think there is a bug in qemu-server (?). How can i help to debug the problem!?

So its a network problem? If so, what network card emulation do you use? Try another one. Also test with other kernels. Does tcpdump give some information?