Is uptime value valid in an ovz container?

camarg

How is this possible?

host node

[root@ovz1 ~]# date && uptime
Thu Jan 24 19:19:41 EET 2013
 19:19:41 up 25 days, 19:58,  1 user,  load average: 0.09, 0.07, 0.08

guest container

root@srv2 [~]#  date && uptime
Thu Jan 24 19:19:42 EET 2013
 19:19:42 up 36 days,  2:20,  1 user,  load average: 0.10, 0.07, 0.06

rds100

Live migrated from another node?

camarg

@rds100 said: Live migrated from another node?

nope

jar

Well you can suspend a container and start it back in it's previous state. I've never considered the possible effect on uptime, but it would make sense.

camarg

@jarland said: suspend a container and start it back in it's previous state

this is not the case either

jar

I'm inclined to call it a bug on that particular container, perhaps template even. Certainly open to being wrong.

camarg

Issued a reboot at the host

host

[root@ovz1 ~]# date && uptime
Fri Jan 25 03:55:31 EET 2013
 03:55:31 up 1 min,  1 user,  load average: 1.05, 0.41, 0.15

guest

root@srv2 [~]# date && uptime
Fri Jan 25 03:55:32 EET 2013
 03:55:32 up 36 days, 10:55,  1 user,  load average: 0.64, 0.26, 0.09

is this normal?

AlexBarakov

Deffinetely not normal. It must be some kind of a bug.

fly

ovz

gl with bug squashing tho

Microlinux

What is the output of 'cat /proc/uptime', is that changing?

Kenshin

Effects from vzctl 4.1 (http://wiki.openvz.org/Download/vzctl/4.1)

etc/init.d/vz: restore running containers after reboot (#781)
etc/init.d/vz: faster restart by doing CT suspend instead of stop (#2325)