Good morning, hope the week is going well for everyone.
We've noted an issue which spans several kernel releases which I
wanted to bounce off the group collective. We are currently observing
the issue all the way from 2.6.32 to 3.4.x so if it is a legitimate
problem it has been around for a long time.
In our spare time we've been working on putting together the mythical
'incremental scrubber' which Neil has been talking about for a long
time. A utility which will start a 'check' run on an MD device and
then capture the completion state after a pre-determined amount of
time so the scrub can be halted and then re-started during the next
evening's session.
When a synchronization or check run of an MD device starts the status
of the check which is found in the following pseudo-file:
/sys/block/mdN/md/sync_completed
Seems to stall and only update at intermittent times while the
/proc/mdstat pseudo-file provides continuous updates.
I'm currently watching a 2.6.32.24 machine which had a RAID5
synchronization started about 25 minutes ago. The contents of the
/proc/mdstat pseudo-file indicate its 6% completed while the
/sys/block/mdN/md/sync_completed file for the device has yet to move
from zero.
I'm also watching a RAID5 synchronization on a 3.4.10 kernel and while
/proc/mdstat is continuously updating the
/sys/block/mdN/md/sync_completed file hasn't budged in about a half
hour.
As I was writing this I just noticed on the 2.6.32.24 machine the
contents of the /sys/block/mdN/md/sync_completed file moved, in one
swoop, from:
0 / 1463746176
To:
91484144 / 1463746176
With respect to providing status information for the re-start of a
check scrub the behavior seems to be anomolous.
Would be interested in experience from others or reflections from Neil.
Have a good afternoon.
Greg
As always,
Dr. G.W. Wettstein, Ph.D. Enjellic Systems Development, LLC.
4206 N. 19th Ave. Specializing in information infra-structure
Fargo, ND 58102 development.
PH: 701-281-1686
FAX: 701-281-3949 EMAIL: greg@xxxxxxxxxxxx
------------------------------------------------------------------------------
"A distributed system is one on which I cannot get any work done because
a machine I have never heard of has crashed."
-- Leslie Lamport
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
