OpenVZ Forum


Home » General » Support » vzctl update breaks classic chkpnt on 2.6.18 based kernel
vzctl update breaks classic chkpnt on 2.6.18 based kernel [message #51334] Wed, 16 April 2014 08:54 Go to next message
Bodysplit is currently offline  Bodysplit
Messages: 20
Registered: July 2007
Junior Member
From: *dip0.t-ipconnect.de
It seems that now released vzctl does not want to checkpoint my CTs anymore:
No ploop support in the kernel, or kernel is way too old.
Make sure you have OpenVZ kernel 042stab058.7 or later running,
and kernel ploop modules loaded.


This seems to be right after the update of vzctl.

Kernel is:
[root@bart ~]# uname -a
Linux bart.netfreax.de 2.6.18-348.4.1.el5.028stab107.2 #1 SMP Mon Jul 1 22:13:01 MSD 2013 x86_64 x86_64 x86_64 GNU/Linux


I have not restarted for a while. It would go to 28stab112.3 if I'm reading grub correctly. Would that fix the problem of vzctl?

Me was only using classic checkpointing non-ploop CTs on Centos 5. Will that not work anymore? I have a testbed running Centos 5 with 2.6.32 kernel as documented in the wiki. Is this the better alternative? I'm still a bit resistent on updating the kernel to .32 although the testbed showed no problems.


--
This doesn't look comfortable...
Re: vzctl update breaks classic chkpnt on 2.6.18 based kernel [message #51351 is a reply to message #51334] Fri, 25 April 2014 14:11 Go to previous messageGo to next message
Paparaciz
Messages: 302
Registered: August 2009
Senior Member
From: *gaumina.lt
kernel: 2.6.18-371.6.1.el5.028stab112.3

I can't reproduce problem

maybe you try to resume ct which was suspended from other server with rhel6 based kernel?

I recommend on rhel5 based server after update to vzctl 4.7 make 2 changes in /etc/vz/vz.conf:

CONFIGFILE="basic"
## Filesystem layout for new CTs: either simfs (default) or ploop
VE_LAYOUT=simfs

Re: vzctl update breaks classic chkpnt on 2.6.18 based kernel [message #51353 is a reply to message #51334] Sat, 26 April 2014 09:55 Go to previous message
Bodysplit is currently offline  Bodysplit
Messages: 20
Registered: July 2007
Junior Member
From: *dip0.t-ipconnect.de
Paparaciz,

it seems you pointed into the right direction! My VE_Layout was set to ploop, which is not available in the EL5 kernel. According to my records, ve_layout was unset before and seems to have been activated with this update.

Still, now I've changed the box to 2.6.32 and ext4. So far running good.


--
This doesn't look comfortable...
Previous Topic: Kernel Panic starting VM
Next Topic: Cron weird behaviour (running twice)
Goto Forum:
  


Current Time: Tue Sep 25 23:38:31 GMT 2018