Got these messages this morning

14/09/2007 6:28:49 AM|rosetta@home|Scheduler RPC succeeded
14/09/2007 6:28:49 AM|rosetta@home|Message from server: Project encountered internal error: shared memory
14/09/2007 6:28:49 AM|rosetta@home|Deferring communication for 1 hr 0 min 0 sec
14/09/2007 6:28:49 AM|rosetta@home|Reason: project is down

---

I'm running v5.78 on a 600 MHz machine. Three of the last nine WUs that I've uploaded reported a "Validate error" (103801478/94222135; 104683771/94998439; and, 104848402/95151762). In total, these three WUs represent just shy of 19 CPU hours, with a combined credit claim of just over 58.

As you might imagine, wasting 19 hours of CPU time because every 1 out of 3 WUs is rejected has me a bit frustrated with R@H! (I'm also running World Community Grid and Seti@Home, neither of which are producing errors.) Is anyone else experiencing similar problems with v5.78?

Here is another one

104430230 94762288 9 Sep

---

Jmarks

I realize that 5.78 is (unpleasant) history now, but as a historical record, and in the hope that this information will make future versions of Rosetta more stable, I offer the following:

Five out of six Capri WUs killed by the watchdog, the casualties were:

WU 95336696
WU 95336686
WU 95336658
WU 95336625
WU 95366073

I won't try to link them all here, but they are all individually linked in my posts in this thread.

The sole survivor of the batch was:

WU 95336673

Even though it was a success, and this is a thread for problems, I have linked it here since there doesn't seem to be a place to post the (rare) successes in v5.78...

This represent about 90 hours on a pretty decent computer, about half of which was accomplished AFTER the problem with 5.78 was identified and (we hope) fixed.

When/If a situation like this arises again, I would specifically request a notice on the home page asking users to ABORT the WUs in question. I suspect the project would be better served thereby.

Respectfully,
David Emigh

---

Rosie, Rosie, she's our gal,
If she can't do it, no one shall!

v5.80 appears to have inherited the same problem (Message 46203)!

WU

**********************************************************************
Rosetta score is stuck or going too long. Watchdog is ending the run!
Stuck at score -457.996 for 900 seconds
**********************************************************************
GZIP SILENT FILE: .xx1he8.out

same problem with this WU

**********************************************************************
Rosetta score is stuck or going too long. Watchdog is ending the run!
Stuck at score 117.804 for 900 seconds
**********************************************************************

thats 3 work units this week that got stuck and gave me only 20 points.

---

this wu got stuck
# cpu_run_time_pref: 21600
# random seed: 1279344
**********************************************************************
Rosetta score is stuck or going too long. Watchdog is ending the run!
Stuck at score -231.574 for 900 seconds
**********************************************************************

yet another 20 instead of actual points, 4 of them now out of over 20 wu's

---

more errors

104452248 94782448 10 Sep 2007 0:41:23 UTC 15 Sep 2007 1:03:19 UTC Over Validate error Done 21,553.84 --- ---
104452246 94782446 10 Sep 2007 0:41:23 UTC 15 Sep 2007 1:03:19 UTC Over Validate error Done 21,609.22 --- ---

---

1he8__BOINC_CAPRI14_DOCK_FIXBACKBONE-1he8_-nodimerloop_plexinmonomer__2067_1639_0 validate error

---