Thanks for the info. I'll see what I can do about the hibernate vs the shut down.
CJSL
Crunching for a better future...

littlepeaks,

Since the invalid prob is only MCM, I've already resorted to planned shutdown/boot, flip profile to short FAHV and just let it rip. Then at some point only FAHV are crunching, then do whatever. After, flip the profile back to include/exclusively do MCM again. No time lost [hardly], no idling cores, maximum crunch return. Been doing same for CEP2 due the wide apart checkpointing... there's no checkpoint tuned planning when 4-8 run concurrent.

Sorry, triple Emming ** is a must these days [or just shrug shoulder and take the who cares beyond the Set-And-forget route. Just crunch and let the fallout be the fallout]

** Micro Management Methodology.

FWIW, I've taken a look thru my Results Status at the MCM WUs that went Pending Verification over the last several days. I looked at the reslts log for those that got Invalid on each WU and the BOINC levels were across multiple BOINC 6x and 7x levels. The one thing they all had was multiple "COmmand line, Initializing, Running" entries. I think I even found invalids with the same size Results.out file. From the various thread on this issue, it would seem this may help confirm the problem is with reloading from checkpoints. I normally run with LAIM on so I have to force a machine reboot to test if that causes the problem for my copy of a WU. Since the techs are working on this, I'm not going to force that but will try to track what happens if I have a reboot soon.

Project Name: Mapping Cancer Markers
Created: 11/30/2013 01:51:21
Name: MCM1_0000308_0969
Minimum Quorum: 2
Replication: 4

Result Name App Version Number Status Sent Time Time Due /
Return Time CPU Time / Elapsed Time (hours) Claimed/ Granted BOINC Credit
MCM1_ 0000308_ 0969_ 3-- 726 Valid 07/12/13 07:00:19 07/12/13 16:50:40 4.74 96.5 / 98.1
MCM1_ 0000308_ 0969_ 2-- 726 Valid 06/12/13 23:51:28 07/12/13 06:59:46 4.80 99.6 / 98.1
MCM1_ 0000308_ 0969_ 1-- 726 Invalid 01/12/13 12:58:24 02/12/13 18:41:09 6.36 126.6 / 49.0
MCM1_ 0000308_ 0969_ 0-- 726 Invalid 01/12/13 12:58:21 06/12/13 23:51:11 15.21 112.5 / 49.0

All have the same Result.out.

_0 restarted five times.
_1 restarted three times.
_2 and _3 ran straight through.

It would be interesting to find one where two Invalids both restarted the same number of times, but I think I'm going to turn this machine over to FAAH until this is fixed. It's just wasting everyone's time to have a machine that gets switched off run MCM1. (And, no, it's not in my control and it's a desktop machine, so there's no option to suspend it.)

Weird.

Before I changed things, this unit went valid:

Project Name: Mapping Cancer Markers
Created: 12/05/2013 23:25:38
Name: MCM1_0000433_1016
Minimum Quorum: 2
Replication: 2

Result Name App Version Number Status Sent Time Time Due /
Return Time CPU Time / Elapsed Time (hours) Claimed/ Granted BOINC Credit
MCM1_ 0000433_ 1016_ 1-- 726 Valid 06/12/13 23:25:28 07/12/13 07:55:35 1.96 73.3 / 62.8
MCM1_ 0000433_ 1016_ 0-- 726 Valid 06/12/13 23:25:25 08/12/13 18:06:27 7.18 52.3 / 62.8

and although the wingman went straight through, the unit on this machine restarted twice - and it still validated.

Maybe something has now changed? Anyway, I'll let it run some more and see what happens.