peletiah / openvz Goto Github PK

When looking at /proc/user_beancounters the held value of physpages is lower than the limit but check_beancounters reports an issue:

uid resource held maxheld barrier limit failcnt
100:
physpages 3641229 3679180 0 4194304 0

/usr/bin/check_beancounters -f
app100: physpages

It seems that your check compares the current value to the limit * 0.9... why?

If there are past beancounter failures for containers that have been destroyed the script throws the following error:

/check_beancounter.py -i -e [email protected]

Traceback (most recent call last):
File "./check_beancounter.py", line 198, in
current_data = compare_data(beancounter_data,data_read,count,email)
File "./check_beancounter.py", line 119, in compare_data
count_change=cntcheck(data_read,current_data,veid,fields,count,count_change)
File "./check_beancounter.py", line 71, in cntcheck
hostname=find_veid(veid)
File "./check_beancounter.py", line 24, in find_veid
veid_conf=open('/etc/vz/conf/' + str(veid) + '.conf','r')
IOError: [Errno 2] No such file or directory: '/etc/vz/conf/319.conf'

I don't know enough python yet, but wrapping the first line in find_veid in a try, to make sure the conf file exists before working with it should solve the problem. It should be safe to ignore any beancounter failures for containers not running or that don't have config files.

Hi,
I use proxmox for running openvz. There is new behaviour in kernel 2.6.32-7-pve: barrier is 0 for oomguarpages of HN.
So now I'm always getting error OpenVZ HN: oomguarpages from nagios because of this.

I don't know if this is new feature openvz or bug. I also reported this in proxmox forum:
http://forum.proxmox.com/threads/8817-barrier-of-oomguarpages-in-2-6-32-7-pve