init-logger chewing CPU...

[John Martin]

I installed a new proxmox environment wednesday this week. Everything OK.

Since I had some problems regarding disk configuration (solved now), I reinstalled today, using the bare-metal ISO for 1.5.

Unfortunately, now I have another problem. The init-logger consume 99% CPU.

Installation details:

# uname- a
Linux pve 2.6.18-2-pve #1 SMP Mon Feb 1 10:45:26 CET 2010 x86_64 GNU/Linux

12:16:59 up 12 min, 1 user, load average: 1.44, 1.08, 0.64
143 processes: 140 sleeping, 3 running, 0 zombie, 0 stopped
CPU0 states: 0.2% user 0.0% system 0.0% nice 0.0% iowait 99.3% idle
CPU1 states: 14.1% user 85.4% system 0.0% nice 0.0% iowait 0.0% idle
CPU2 states: 0.2% user 0.0% system 0.0% nice 0.0% iowait 99.3% idle
CPU3 states: 0.1% user 0.0% system 0.0% nice 0.0% iowait 99.4% idle
Mem: 6083488k av, 658824k used, 5424664k free, 0k shrd, 16796k buff
361852k active, 213772k inactive
Swap: 5242872k av, 0k used, 5242872k free 329648k cached

PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND
6275 root 25 0 92 12 4 R 99.9 0.0 11:46 0 init-logger

dpkg-query -l | grep pve
ii libpve-storage-perl 1.0-10 Proxmox VE storage management library
ii pve-firmware 1.0-3 Binary firmware code for the pve-kernel
ii pve-kernel-2.6.18-2-pve 2.6.18-5 The Proxmox PVE Kernel Image
ii pve-manager 1.5-8 The Proxmox Virtul Environment
ii pve-qemu-kvm-2.6.18 0.9.1-5 Full virtualization on x86 hardware
ii vzctl 3.0.23-1pve8 OpenVZ - server virtualization solution - co

The container is based on ubuntu-8.04-3 and running ebox-office.


John Martin

New information:

I have now reprodused the error with the original template: (Ubuntu Hardy (standard) ubuntu-8.04-standard_8.04-3_i386.tar.gz) The error occures both for venet and vmbr networks.​

 

[dietmar]

I have now reprodused the error with the original template: (Ubuntu Hardy (standard) ubuntu-8.04-standard_8.04-3_i386.tar.gz) The error occures both for venet and vmbr networks.

How to repruduce it - exactly?

 

[John Martin]

How to repruduce it - exactly?

I installed PVE from scratch, applied all updates to the PVE-server with aptitude. Both, my own modified templates and the original Ubuntu-8.04 template had a problem with init-logger.

Now I have installed from scratch again, no updates applied. The original templates have no problems. My modified templates have the problem and I must investigate this myself.

The only difference I find i PVE-packages are:

ii libpve-storage-perl 1.0-8
ii libpve-storage-perl 1.0-10

and

ii pve-manager 1.5-5
ii pve-manager 1.5-8

Reason:
The /var/log/init.fifo was packed into the template. When generating a new machine, there will be an empty file with the same name. This file is not deleted when the init.fifo is generated. Therefore, the fifo will not be created the proper way.
I will now try to upgrade the PVE, and see if the problem is still present.

 

[Kalin Bogatzevski]

exactly the same problem with me here
i deleted /var/log/init.* and rebooted the container. the init.fifo and init.log got recreated and everything's now fine.
thanks for the hint!

 

[mbm100]

Hello,

I noticed this same behavior, but only when using NFS storage.

I solved this issue with so.

Inside the CT who has 100% init-logger CPU usage, run


# chmod 770 /var/log/init.fifo
# chmod 770 /var/log/init.log
# reboot