Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Advertise on LowEndTalk.com
OpenVZ Kernel Problems with 2.6.18-308.8.2.el5.028stab101.x ?
New on LowEndTalk? Please read our 'Community Rules' by clicking on it in the right menu!

OpenVZ Kernel Problems with 2.6.18-308.8.2.el5.028stab101.x ?

KrisKris Member
edited April 2013 in General

Anyone have problems with these kernels on OpenVZ : 2.6.18-308.8.2.el5.028stab101.3 / 2.6.18-308.8.2.el5.028stab101.1

Specifically, numerous kernel panics that erupted out of nowhere, 3 days after upgrading the kernel.

Data center's tested the memory, every thing looks OK from MegaLSI, but random kernel panics ensue.

They've blamed these OpenVZ kernels. As soon as 2.6.18-348.3.1.el5.028stab106.2 was released, we updated - no issues.

What I dont understand is we were running 028stab101.1 for 40+ days before upgrading to 101.3, no issues.

I still suspect a lingering RAM issue that will probably show it's face, but apparently, it was those OpenVZ kernels.

Same config as everyone in Buffalo has.

Brought down machine and fsck -yf /vz to make sure no disk issues, and checked all RAID integrity data, no errors.

Comments

  • 101.3 was bad for us when upgrading - kernel panics, tun/tap related. The fact that the new vzctl / vztools suspends and snapshots the VPSes and then restores them on reboot is the reason. If the VPS is just stopped and then restarted on the new kernel - no problems.

    -

  • RalliasRallias Member, Provider

    @rds100 said: 101.3 was bad for us when upgrading - kernel panics, tun/tap related. The fact that the new vzctl / vztools suspends and snapshots the VPSes and then restores them on reboot is the reason. If the VPS is just stopped and then restarted on the new kernel - no problems.

    This... we had a few issues with KP's that showed tun/tap on screen. An upgrade to the 104.* series fixed these.

  • KrisKris Member
    edited April 2013

    We were getting fatal sync exceptions, nothing else in terms of messages - keep in mind the machines were power-cut incorrectly by the data center - incorrectly marked PDU's - technician hard power-cycled during spin-ups (twice)

    Reverted to 101.1 and still got kernel panics. (which was stable for weeks)

    No messages in dmesg or anything except Fatal sync exception. After hours of gawking we had them test the memory, of course claimed nothing was wrong - but then magically came up fine on 028stab106.2, after the machine was down and "they didn't replace the RAM"

Sign In or Register to comment.