FYI, a possible fix for a bug I was seeing with software RAID on
2.6.26+OpenVZ a while ago, (and reported here at the time) has been
highlighted by Ben Hutchings.
The bug is still present in current production kernels (incl RHEL6
etc.). If I was indeed seeing this bug, then other changes between
2.6.26 and 2.6.32 meant that the bug was no longer frequently triggered
on my test environment.
I'm posting here in case OpenVZ continues to tickle this bug more
frequently than non-OpenVZ installs for some reason - Ben's email with
patch is here: