Now I definitely know that I was using the wrong bait.

Not an expert in Boinc stuff but here it goes:

unit1: core2duo 2.66g, 4gig ram, Win7 64-bit: 2 WU done without a hickup

unit2: celeron (C2D) 2.50g, 2g ram, WinXP x86: hard drive activity was high, 2 WU unit done ok

unit3: AMD64 2cores, 2.0ghz, 1g ram, WinXP x86: The 2 WU went over the 12hrs mark, everything else was OK.

Sidenote: 75% memory was allocated on all units and hard drive allocation was over 20gb for Boinc.

Overall no problems, however on one of my machines I received the following set of messages twice (2nd set one hour later, since then no additional ones):

11/2/2010 7:27:56 AM World Community Grid Task BETA_E200520_233_A.24.C21H14OSSe.25.3.set1d06_1 exited with zero status but no 'finished' file
11/2/2010 7:27:56 AM World Community Grid If this happens repeatedly you may need to reset the project.
11/2/2010 7:27:56 AM World Community Grid Task BETA_E200520_338_A.25.C21H14N2S2.441.0.set1d06_0 exited with zero status but no 'finished' file
11/2/2010 7:27:56 AM World Community Grid If this happens repeatedly you may need to reset the project.
11/2/2010 7:27:56 AM World Community Grid Task BETA_E200520_339_A.25.C22H14OS2.41.4.set1d06_1 exited with zero status but no 'finished' file
11/2/2010 7:27:56 AM World Community Grid If this happens repeatedly you may need to reset the project.
11/2/2010 7:27:56 AM World Community Grid Task BETA_E200520_797_A.25.C22H15NOS.112.2.set1d06_0 exited with zero status but no 'finished' file
11/2/2010 7:27:56 AM World Community Grid If this happens repeatedly you may need to reset the project.
11/2/2010 7:27:56 AM World Community Grid Task BETA_E200520_793_A.25.C21H15N3S.136.2.set1d06_0 exited with zero status but no 'finished' file
11/2/2010 7:27:56 AM World Community Grid If this happens repeatedly you may need to reset the project.
11/2/2010 7:27:56 AM World Community Grid Task BETA_E200520_806_A.25.C23H15NSi.67.4.set1d06_0 exited with zero status but no 'finished' file
11/2/2010 7:27:56 AM World Community Grid If this happens repeatedly you may need to reset the project.
11/2/2010 7:27:56 AM World Community Grid Task BETA_E200520_808_A.25.C22H15NOS.16.1.set1d06_1 exited with zero status but no 'finished' file
11/2/2010 7:27:56 AM World Community Grid If this happens repeatedly you may need to reset the project.
11/2/2010 7:27:56 AM World Community Grid Task BETA_E200520_815_A.25.C21H15NO2Si.27.4.set1d06_0 exited with zero status but no 'finished' file
11/2/2010 7:27:56 AM World Community Grid If this happens repeatedly you may need to reset the project.

Picked up 9 Beta tasks, 3 each on 3 machines; a Q6600 (W7 x64), a Phenom II 940 (Vista x64) and an i7-920 (XP x86). 1 Valid (Phenom), 6 PV and 2 IP on the Phenom.

The tasks seem sound enough, but the efficiency is still poor:
CPU Time 5h 37min, Elapsed Time 6h 43min – 83.6% efficiency.
2 tasks IP have 4 image names (Task Manager):
wcgrid_beta11_6.35_windows_intelx86 details,
I/O Reads, Writes and Other; 827919, 56850, 73951
I/O Reads (Bytes), I/O Write (Bytes), I/O Other (Bytes); 203025919, 517148514, 730160

This is my (8h) Valid Task:
Project Name: Beta - The Clean Energy Project - Phase 2 Created: 01/11/10 Name: BETA_E200520_690_A.25.C22H15NOS.10.3.set1d06 Minimum Quorum: 2 Replication: 2
Result Name App Version Number Status Sent Time Time Due /
Return Time CPU Time (hours) Claimed/ Granted BOINC Credit BETA_ E200520_ 690_ A.25.C22H15NOS.10.3.set1d06_ 1-- 635 Valid 02/11/10 00:00:57 02/11/10 03:04:24 1.73 33.3 / 31.6 BETA_ E200520_ 690_ A.25.C22H15NOS.10.3.set1d06_ 0-- 635 Valid 02/11/10 00:00:49 02/11/10 09:21:47 8.04 159.9 / 168.6
and the details:
Result Name: BETA_ E200520_ 690_ A.25.C22H15NOS.10.3.set1d06_ 0--

<core_client_version>6.10.58</core_client_version>
<![CDATA[
<stderr_txt>
INFO: No state to restore. Start from the beginning.
[00:06:29] Number of jobs = 16
[00:06:29] Starting job 0,CPU time has been restored to 0.000000.
[00:09:00] Finished Job #0
[00:09:00] Starting job 1,CPU time has been restored to 135.798870.
[00:15:26] Finished Job #1
[00:15:26] Starting job 2,CPU time has been restored to 443.386042.
[03:10:19] Finished Job #2
[03:10:19] Starting job 3,CPU time has been restored to 9966.030284.
[03:17:15] Finished Job #3
[03:17:15] Starting job 4,CPU time has been restored to 10314.645719.
[03:22:32] Finished Job #4
[03:22:32] Starting job 5,CPU time has been restored to 10591.594294.
[03:28:02] Finished Job #5
[03:28:02] Starting job 6,CPU time has been restored to 10875.157312.
[03:33:16] Finished Job #6
[03:33:16] Starting job 7,CPU time has been restored to 11152.355489.
[03:40:43] Finished Job #7
[03:40:43] Starting job 8,CPU time has been restored to 11544.932005.
[03:45:28] Finished Job #8
[03:45:28] Starting job 9,CPU time has been restored to 11796.920421.
[03:50:47] Finished Job #9
[03:50:47] Starting job 10,CPU time has been restored to 12076.474213.
[04:03:36] Finished Job #10
[04:03:36] Starting job 11,CPU time has been restored to 12761.490204.
[04:10:48] Finished Job #11
[04:10:48] Starting job 12,CPU time has been restored to 13132.132980.
[04:49:56] Finished Job #12
[04:49:56] Starting job 13,CPU time has been restored to 15212.624716.
[06:14:08] Finished Job #13
[06:14:08] Starting job 14,CPU time has been restored to 19732.753691.
[07:33:58] Finished Job #14
[07:33:58] Starting job 15,CPU time has been restored to 24115.476985.
[09:15:56] Finished Job #15
09:16:22 (376): called boinc_finish

</stderr_txt>
]]>

This is the Wingmans Valid result:

Result Log

Result Name: BETA_ E200520_ 690_ A.25.C22H15NOS.10.3.set1d06_ 1--

<core_client_version>6.10.58</core_client_version>
<![CDATA[
<stderr_txt>
INFO: No state to restore. Start from the beginning.
[02:07:13] Number of jobs = 16
[02:07:13] Starting job 0,CPU time has been restored to 0.000000.
[02:10:07] Finished Job #0
[02:10:07] Starting job 1,CPU time has been restored to 153.734375.
[02:16:49] Finished Job #1
[02:16:49] Starting job 2,CPU time has been restored to 530.500000.
Application exited with RC = 0xc0000005
[04:01:05] Finished Job #2
[04:01:05] Starting job 3,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #3
[04:01:05] Starting job 4,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #4
[04:01:05] Starting job 5,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #5
[04:01:05] Starting job 6,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #6
[04:01:05] Starting job 7,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #7
[04:01:05] Starting job 8,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #8
[04:01:05] Starting job 9,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #9
[04:01:05] Starting job 10,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #10
[04:01:05] Starting job 11,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #11
[04:01:05] Starting job 12,CPU time has been restored to 6212.562500.
[04:01:05] Skipping Job #12
[04:01:05] Starting job 13,CPU time has been restored to 6212.562500.
[04:01:06] Skipping Job #13
[04:01:06] Starting job 14,CPU time has been restored to 6212.562500.
[04:01:06] Skipping Job #14
[04:01:06] Starting job 15,CPU time has been restored to 6212.562500.
[04:01:06] Skipping Job #15
04:01:07 (2912): called boinc_finish

</stderr_txt>
]]>

I take it you are implementing partial task completion validation :)
Will this work for any error output or just some?

And? The number is too small to waste days of bug hunting and the fix works... those that open the detail see the "Try validation" and those that don't will have these validator slipped tasks forced through the validator when the original deadline passes, the first check as the system wants to know at that time why the quorum is not complete and schedule a repair job. Then it finds it need not.

If the initial validator flagging fails, the "Try validation" will not show until the first midnight run after quorum is complete.

There was a plan of adding a bit of code that if a member would open up the detail and a "Try Validation" flag is showing, to take that as sign to push the button and automatically move it into the queue i.e. a "scheduled for validation" appearing automatically. That was thought to be very low load. Running a script to pick up those that missed the boat initially was causing too much burden and thus not put in place.

Just repeating what I've read before.