No log issued on openvz resource shortage events

[philten]

Hello,

According to what is reported on many web threads dedicated to openvz,
resource shortage issues should be reported in by an increased failcnt
visible in /proc/user_beancounters and a message in dmesg, similar to:

Fatal resource shortage: privvmpages, UB 102.

On my proxmox 1.4 server, I created a privvmpages shortage on purpose
(by lowering privvmpages values)

I do get the failcnt output with cat /proc/user_beancounters:

pve1:/var/log# /usr/sbin/vzctl exec 104 cat /proc/user_beancounters
Version: 2.5
uid resource held maxheld barrier limit failcnt
104: kmemsize 345261 1488046 9223372036854775807 9223372036854775807 0
lockedpages 0 0 131072 131072 0
privvmpages 419 1195 1144 1644 23 <<<<<<<< !!!!!
shmpages 0 0 9223372036854775807 9223372036854775807 0

but *nothing* reported in dmesg. Furthermore I do not have any logs regarding openvz CT events such as start or stop.

This is very surprising to me as this seem basic openvz log reports.
Since these logs are important in case of CT issues
I started this thread hoping to get more information.

Since failcnt counters are reset after CT restart, it would helpfull to find these logs.

Is this a setting matter such as a log level somewhere ? or can it be found in other log files ?

Background info:

pve1:/var/log# pveversion -v
pve-manager: 1.4-9 (pve-manager/1.4/4390)
qemu-server: 1.1-6
pve-kernel: 2.6.24-16
pve-qemu-kvm: 0.11.0-2
pve-firmware: 1
vncterm: 0.9-2
vzctl: 3.0.23-1pve3
vzdump: 1.2-3
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1

pve1:/var/log# uname -a
Linux pve1 2.6.24-8-pve #1 SMP PREEMPT Fri Oct 16 11:17:55 CEST 2009 x86_64 GNU/Linux

I noticed the same behavior on a proxmox server 1.5

I started suspecting this fact while troubleshooting a CT which obviously had memory shortage
but nothing reported in HN dmesg or syslog.

Thanks for all information.

Phil Ten

 

[dietmar]

Strange, I get those logse here

# vzctl stop 104
# vzctl start 104

generates the following logs:

Feb 12 06:29:25 oahu kernel: CT: 104: stopped
Feb 12 06:29:38 oahu kernel: CT: 104: started

 

[philten]

I just tried again with a low privvmpages value

I do have the same as you in syslog, but nothing in dmesg,
and no "Fatal resource shortage" listed neither in syslog nor dmesg.

pve1:/var/log# grep 104 syslog
Feb 12 09:32:11 pve1 pvedaemon[6982]: restarting VM 104 on node 0 (localhost)
Feb 12 09:32:20 pve1 kernel: CT: 104: stopped
Feb 12 09:32:21 pve1 kernel: CT: 104: started
Feb 12 09:32:23 pve1 pvedaemon[6982]: VM 104 restarted

pve1:/var/log# /usr/sbin/vzctl exec 104 cat /proc/user_beancounters
Version: 2.5
uid resource held maxheld barrier limit failcnt
104: kmemsize 353658 1660102 9223372036854775807 9223372036854775807 0
lockedpages 0 0 131072 131072 0
privvmpages 419 1175 1144 1644 18 <<<<<<<<< resource shortage
shmpages 0 0 9223372036854775807 9223372036854775807 0



Phil Ten

 

[philten]

I add that I seen this symptom on several HN. In fact, I never seen these resource
shortage messages on any Proxmox server.

Would anybody have an idea why/where these kernel messages would be blocked ?

I double checked rsyslog.conf but found nothing there.
To be sure I redirected *.* to the console and created an openvz resource shortage.
As expected, I got CT start/stop messages, but no resource shortage logs
Very strange ???
Phil Ten