Proxmox VE 1.9 released!

[ciairo]

Hi there!

first at all i want to thank the Proxmox team for the good work! im proud to be using it!

i have 5 towers running... and before upgrading them all i tried with one of them first and i am having a strange situation... i have no idea if this is do to the upgrade, but since it's only happening with the VM's on the 1.9 environment it makes me think so...

the problem is:
starting the server after a power down everything works perfectly - but after a while all of the NIC's that are on my VMBR1 stop working... if i do a "ifconfig" inside the VM i see my NIC up&running but if i ping the broadcast addy only myself responds... restarting the network services on the VM get's the NIC running again, but not for long.

i did all the testing a could, all linux vm's seems to be working perfectly... i checked all the interconnections as well... does any body have any hints?

details:
the VM's in question are all RedHat with 2 Nic's one of them (VMBR0) works perfectly, the other one (VMBR1) does not... the only output i could find is:

" Oct 6 08:44:55 HOSTNAME kernel: NETDEV WATCHDOG: eth0: transmit timed out "
" Oct 6 08:44:55 HOSTNAME kernel: e1000: eth0: e1000_watchtog: NIC Link is Up 1000 Mbits Full Duplex "

to me it seems like the NIC is up - but looses the connection at the other end... physically (switch, cables etc) are fine.

any hints? or idea's?

thank you...

ciairo.

 

[tom]

post details about your VM (VMID.conf), what OS do you run? if you run virtio net, make sure you use the latest drivers.

 

[ciairo]

Dear Tom,

i tried looking for the VMID.conf file but i could not find it...

the VM's are:

RedHat EL 3 with a 2.6.21-37 Kernel
1 CPU Socket
1 CPU Core
1024MB of Ram
IDE
50GB of *.raw disk
2x Intel e1000 NIC's on different VMBR's (one works the other gives me trouble)

anything else ?

thanks again,

ciairo

 

[tom]

Dear Tom,

i tried looking for the VMID.conf file but i could not find it...

see /etc/qemu-server/VMID.conf

the VM's are:

RedHat EL 3 with a 2.6.21-37 Kernel
1 CPU Socket
1 CPU Core
1024MB of Ram
IDE
50GB of *.raw disk
2x Intel e1000 NIC's on different VMBR's (one works the other gives me trouble)

anything else ?

thanks again,

ciairo

quite old, I have no experience with these system.

 

[jleg]

Dear Tom,

i tried looking for the VMID.conf file but i could not find it...

the VM's are:

RedHat EL 3 with a 2.6.21-37 Kernel
1 CPU Socket
1 CPU Core
1024MB of Ram
IDE
50GB of *.raw disk
2x Intel e1000 NIC's on different VMBR's (one works the other gives me trouble)

anything else ?

thanks again,

ciairo

...are you sure that the *physical* NIC (which that 'failing' vmbr is pointing to) still is operable when your VM's NIC failed?
It's just because we had some troubles with RHEL6-kernel, which Proxmox is based on with current release.
In case your physical NIC is also affected, it may be worth updating the driver to latest - this solved our issues for e1000e and igb...

 

[tom]

our latest kernel uses the very latest e1000e (1.6.2-NAPI ) and also igb (3.2.9). so which driver did you upgrade, do which version?

 

[jleg]

our latest kernel uses the very latest e1000e (1.6.2-NAPI ) and also igb (3.2.9). so which driver did you upgrade, do which version?

ok, wasn't aware of that - i referred to RHEL6 (which runs 1.2.7/1.2.20 for e1000e)...
though it's imo still interesting to know whether physical NIC is affected...

 

[carligraph]

That is not necessary.

but the module didn't load at boot time without that .. !?

 

[jmrivera]

i've tryed to install but display this error :

Disk error 80, AX = 4284, drive 9F

any idea

 

[jleg]

i've tryed to install but display this error :



any idea

burn a new CD, use another burner, use slower burn speed...

 

[ciairo]

Hey there!

after fighting with this issue for over 2 weeks i had to give up! there was nothing a could do to fix this network problem! so i had to backup - re-install the 1.8 version and restore everything! now the 1.8 has been up & runnung for 5 days with no problem what so ever! i did not change nothing on the VM's and alternated there settings.. so it is obvious that this issue was something with the 1.9 version is not right, i am using ProxMox in a professional production and want to keep using it - but i need some help on this one, i can't upgrade until then....

any ideas?

thank you - and wish you all a great start of the week.

 

[linum]

any ideas?

I'm sorry but I didn't have something really useful but I think your thoughts about "something with version 1.9" seems right to mee. Did you take a look at the thread http://forum.proxmox.com/threads/7330-VirtIO-Win2003-warnings-errors-reboots ? I can't see what kind off network drivers you're using but if you're running virtio drivers maybe you have the same issue like the one I've described at http://forum.proxmox.com/threads/7318-WinXP-virtio-proxmox-1.9-gt-crash.

 

[ciairo]

no, unfortunately this is not our problem! i also doubt that this error is a problem with the NIC, i tried the e1000 and the rtl8139 for kick's to see if it was the NIC. + i am running Linux machines not WIN.

my wild guess is that the 1.9 uses a kernel with a different version of Watchdog and after a not static interval sending the NIC's in timeout - it is really strange! because it occurs random and only on 1 typology of VMs... like i have mentions it drove me insane until the point of restoring a 1.8 environment - the real prob is that i can't upgrade until i solve this issue...

any other hints?

Ciairo

 

[ciairo]

Hi there! i'm sorry if i am still nagging! but i got news!

i have a server where i have the 1.9 to do testing for this issue! and listen to this... if i run the 1.9 using the older kernel (2.6.32-4.pve) my problem is gone! so i have to ask the person who patched the kernel to 2.6.32-6 what did he touch that could create this network issue Watchdog? Network functions? if he could send me both .config's id like to adapted and patch it myself, i understand that this option could not be possible - so in that case please work with me here

thank you

Ciairo

 

[ciairo]

Hi there! i'm sorry if i am still nagging! but i got news!

i have a server where i have the 1.9 to do testing for this issue! and listen to this... if i run the 1.9 using the older kernel (2.6.32-4.pve) my problem is gone! so i have to ask the person who patched the kernel to 2.6.32-6 what did he touch that could create this network issue Watchdog? Network functions? if he could send me both .config's id like to adapted and patch it myself, i understand that this option could not be possible - so in that case please work with me here

thank you

Ciairo

Please!

Thank you,
Ciairo

 

[dietmar]

so i have to ask the person who patched the kernel to 2.6.32-6 what did he touch that could create this network issue Watchdog? Network functions?

2.6.32-4 is based on debian squeeze kernel
2.6.32-6 is based on RHEL61 kernel

You can find the config in the /boot/ directory.

 

[gkovacs]

Hi there! i'm sorry if i am still nagging! but i got news!

i have a server where i have the 1.9 to do testing for this issue! and listen to this... if i run the 1.9 using the older kernel (2.6.32-4.pve) my problem is gone! so i have to ask the person who patched the kernel to 2.6.32-6 what did he touch that could create this network issue Watchdog?

Ciaro, your issue is most likely related to the kernel panic we've been experiencing with 2.6.32-6:
http://forum.proxmox.com/threads/7118-Kernel-panic-with-2.6.32-6-and-multi-cpu-OpenVZ

According to the OpenVZ team, the bug has been found and the kernel patched:
http://bugzilla.openvz.org/show_bug.cgi?id=2041

Hopefully the Proxmox devs will release a stable kernel with that patch soon, it's already in testing, called 2.6.32-6-pve_2.6.32-49.

 

[ciairo]

Good Morning!

thank you for your post! !!! - yesterday i spend the hole day diff'ing the two .config's and trying to see if i can compile a kernel that fit's i was concentrating on the networking part - reading the posts posted and seeing the bugzilla post i'll have to re-look at everything!

we did not get any panics from this kernel! and we have several servers that are running 24/7 - besides of the networking issue the 1.9 with the 2.6.32-6 runs stable.... if you like you could post me the new release in test and i will try it out to see if that patch help's also in this situation.

thanks again!! + keep up the good work.

Ciairo

 

[blender]

Hi,
for some reason i don't understand KSM is not working for me.

Did you ever find out what was wrong? I have the same issue on two hosts I recently upgraded to 1.9.

host1:

host1:~# pveversion -v
pve-manager: 1.9-26 (pve-manager/1.9/6567)
running kernel: 2.6.32-6-pve
proxmox-ve-2.6.32: 1.9-50
pve-kernel-2.6.32-6-pve: 2.6.32-50
qemu-server: 1.1-32
pve-firmware: 1.0-14
libpve-storage-perl: 1.0-19
vncterm: 0.9-2
vzctl: 3.0.29-3pve1
vzdump: 1.2-16
vzprocps: 2.0.11-2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.15.0-1
ksm-control-daemon: 1.0-6

host1:~# free
total used free shared buffers cached
Mem: 12263176 6756768 5506408 0 269912 189720
-/+ buffers/cache: 6297136 5966040
Swap: 11534328 0 11534328

host1:~# cat /sys/kernel/mm/ksm/pages_sharing
0

host2:

host2:~# pveversion -v
pve-manager: 1.9-26 (pve-manager/1.9/6567)
running kernel: 2.6.32-6-pve
proxmox-ve-2.6.32: 1.9-50
pve-kernel-2.6.32-6-pve: 2.6.32-50
qemu-server: 1.1-32
pve-firmware: 1.0-14
libpve-storage-perl: 1.0-19
vncterm: 0.9-2
vzctl: 3.0.29-3pve1
vzdump: 1.2-16
vzprocps: 2.0.11-2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.15.0-1
ksm-control-daemon: 1.0-6




host2:~# free
total used free shared buffers cached
Mem: 8105780 6491120 1614660 0 196584 1985508
-/+ buffers/cache: 4309028 3796752
Swap: 4194296 0 4194296

host2:~# cat /sys/kernel/mm/ksm/pages_sharing
0

Both hosts have CentOS/RHEL guests.

A third host that I haven't updated yet and is running 2.6.35 kernel also with CentOS/RHEL guests has 280,000 shared pages.

Please advise.

Thanks!

 

[abkrim]

After a simple apt-get update; apt-get upgrade

Configurando vzctl (3.0.29-3pve1) ...Starting OpenVZ: failed to load module vzmon..failed
invoke-rc.d: initscript vz, action "start" failed.
dpkg: error al procesar vzctl (--configure):
el subproceso post-installation script devolvió el código de salida de error 1
dpkg: problemas de dependencias impiden la configuración de pve-manager:
pve-manager depende de vzctl (>= 3.0.23-1pve2); sin embargo:
El paquete `vzctl' no está configurado todavía.
dpkg: error al procesar pve-manager (--configure):
problemas de dependencias - se deja sin configurar
Se encontraron errores al procesar:
vzctl
pve-manager
E: Sub-process /usr/bin/dpkg returned an error code (1)
root@ns383717:~# pveversion -v
pve-manager: not correctly installed (pve-manager/1.9/6567)
running kernel: 2.6.32-6-pve
pve-kernel-2.6.32-6-pve: 2.6.32-50
qemu-server: 1.1-32
pve-firmware: 1.0-14
libpve-storage-perl: 1.0-19
vncterm: 0.9-2
vzctl: not correctly installed
vzdump: 1.2-16
vzprocps: 2.0.11-2
vzquota: 3.0.11-1dso1

IDEAS?