Critical ZFS bug 6214 in L2Arc module

[gea]

There is a critical bug in current OpenZFS in L2Arc cache modul
that can lead to data corruption.

affected distributions: current OpenZFS

suggested action:
- remove L2Arc
- check pool (scrub, dbd)


http://echelog.com/logs/browse/illumos/
http://blog.gmane.org/gmane.os.omnios.general/day=20150912

 

[dave_the_nerd]

Interesting. Don't use an L2ARC in my rig, but thanks for the heads-up.

 

[gea]

update

Bug is fixed, check for OS updates for systems that includes newest ZFS L2ARC enhancements, ex OmniOS 151014, OI Hipster, SmartOS etc

see illumos . org /issues/6214

 

[XavierMace]

Does this affect Solaris 11.2 or just the forks?

 

[gea]

The problem affects current Illumos (Upstream for OpenZFS) and its improvements to reduce RAM needs of L2Arc.

Oracle Solaris or older ZFS distributions that do not include this improvements are not affected.

 

[CiPHER]

For FreeBSD, this bug was only present in -CURRENT and never reached any -RELEASE neither did it hit -STABLE. It also does not affect any ZFS setup that did not utilise L2ARC.

 

[gea]

Update

A fix is available for OmniOS 151014 at
http://omnios.omniti.com/wiki.php/ReleaseNotes/r151014
Please update (requires a reboot)

more: about how to handle the problem:
http://blog.gmane.org/gmane.os.omnios.general/day=20150914
(read comments, comment at OmniOS discuss if you have additional insights)

 

[CiPHER]

Gea: did this corruption bug hit OmniOS releases, or just master/development branch as was the case for FreeBSD?

 

[gea]

Gea: did this corruption bug hit OmniOS releases, or just master/development branch as was the case for FreeBSD?

The Bug was not discovered in last OmniOS bloody and made its way to the current OmniOS 151014 stable Long Term Service April 2015 edition or the commercially used SmartOS.

http://omnios.omniti.com/wiki.php/ReleaseNotes/r151014 in feature
"Reduction in RAM usage of ZFS cache devices. "

Tthis is a commercially used production release.
I am glad that OmniTi only need a good weekend fo offer a fix for that bug with the ability to trash a whole pool from now to then in the rare but worst case.

 

[CiPHER]

Ouch!

 

[gea]

I have updated the preconfigured ESXi VM at https://napp-it.org/downloads/napp-in-one_en.html
The new ova template 15d includes the L2Arc fix and TLS modules for Googlemail