While the 2.6.32-696.23.1.el6.x86_64 is an old kernel, it is actually the current kernel for CentOS/RHEL 6. Red Hat has backported a lot of fixes to it and CentOS/RHEL 6 is a distribution that is still under maintenance.

It is still out there in a lot of places and has been in production for a while. That it is old, is a feature, it also means that it is incredibly stable.

Also note that the problem is also reported on a CentOS 7.4 kernel a few posts down.

We used to support Linux 2.6.16, but support for that was dropped a long time ago. Then we had support for 2.6.26, but support was dropped for that too. 2.6.32 support will be maintained until at least 2020 if I recall correctly. I could see us maintaining it well past that too.

Indeed, but we are just end consumers of RHEL's release cycles. In large commercial environment (banking in this case) we are bound to use the supported versions from commercial vendors by law.

My scenario is simple - lots of very stable zfs 0.7.7 running on RHEL 7.4. We saw the 0.7.7 media reports of the bug and upgraded to 0.7.8 in fear of a catastrophe - but all hell broke loose after upgrading. Downgraded to 0.7.7 but got the new 7.5 RHEL kernel and still everything is a mess. Rebuilt some test systems with RHEL release 7.5 and 0.7.7 and still cannot even list a mount point of a brand new zpool without even creating any files on it.

Now we are seriously worried. Can't go back to 0.7.6 easily and any way zfs 0.7.6 modules don't load into RHEL 7.5 release - not compatible. We may have to go back to RHEL 7.4 and 0.7.7 which is still stable on many systems.

You guys are all heroes for your phenomenal work. Please keep to going...

My thoughts exactly, while the RHEL kernel has lots of backported patches - it's by no means complete or near current.