Is the rosetta process still present if you Exit BOINC (not just stopping).
If there is no BOINC.exe process present but a rosetta process then there is something wrong with BOINC I think, since it should kill all child processes when it exits.

Sure. I'll better tell the Boinc devs too. I was already thinking about doing this.
But it is only different Boinc behaviour when the problem arises, it still begins with rosetta getting stuck.

Peter

Watch dog on alert.

https://boinc.bakerlab.org/rosetta/result.php?resultid=36825949

Anders n

---

lol... i see all my posts have been removed.

---

lol... i see all my posts have been removed.

I don't think so. It's just that this thread is getting so long that you have to click for more. Look for a link after the original post from Rhiju that says "Click here to also display the remaining posts."

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

ERROR:: Exit at: .dock_structure.cc line:401

https://boinc.bakerlab.org/rosetta/result.php?resultid=37385834

Anders n

---

ERROR:: Exit at: .dock_structure.cc line:401

https://boinc.bakerlab.org/rosetta/result.php?resultid=37385834

Anders n

Same with:
https://boinc.bakerlab.org/rosetta/result.php?resultid=37283001

I only noticed this as it was the first unit (prematurely) returned by my "new" machine...

--
Mats

---

9/16/2006 2:36:48 PM|rosetta@home|Unrecoverable error for result 1fvk_1_CASPR_1_1fvk_1_yyidrenum_16IGNORE_THE_REST_0001_1224_4988_0 (aborted by user)

Took a slot for a Long time with the reported cpu time hung around 3:18:21, but still merrily sucking electricity.

Result id
Work Unit

---

dag
--Finding aliens is cool, but understanding the structure of proteins is useful.

I have had two stuck WU today, having not seen this issue for months.

Both are on Rosetta v5.25

This one stuck at 78.94%, elapsed time stopped increasing. Running on a single cpu Linux box, BOINC v5.27

That one stuck at 100% and still shown as running. This is on a 2cpu linux box, shared with 2 CPDN WU, BOINC v5.28

Both tasks shown as taking no cpu in top. Both boxes rebooted. Sorry, I did not think to save the files before reboot.

Both tasks then recovered normally, the 100% one going on to report immediately, and the 78.94% one carrying on from its last checkpoint at 78.90% I am pretty sure I wrote down the figures correctly, and it backed off just 0.04% which would mean it got stuck right at the end of a very short decoy. Will let you know if it runs to completion OK or gets stuck again on the way.

Am planning now to update both boxes to latest BOINC Linux client in case that is the issue, though seems unlucky if it is when the boxes are running different BOINC versions.

This is not a complaint, just thought you might like to know.

River~~

---

I have had two stuck WU today, having not seen this issue for months.
Both are on Rosetta v5.25
[...]
Am planning now to update both boxes to latest BOINC Linux client in case that is the issue, though seems unlucky if it is when the boxes are running different BOINC versions.

This seems not to be the issue. They are being stuck somewhere just after having checkpointed, therefore the "insignificant back-off of just 0.04%".

If, then only "Leave apps in memory" could help you. But the "stuck at 100%" will happen anyway - the problem is checkpointing at the 100% instead of finishing (thus able to immediately report when getting its turn later again).

(Have been testing it extensively with debug version of Rosetta 5.25 on Boinc 5.3.31 and now running on Boinc 5.5.15 - it behaves the same with "Leave apps in memory" option set - no stuck problems except the 100% one.)

Peter

I have had two stuck WU today, having not seen this issue for months.
Both are on Rosetta v5.25

The one that stuck before the end ran through the place where it stuck before. Unless I post again here, please assume it went on to finish OK - you'll be able to see it once it uploads anyway.

[...]
Am planning now to update both boxes to latest BOINC Linux client in case that is the issue

This seems not to be the issue. They are being stuck somewhere just after having checkpointed, therefore the "insignificant back-off of just 0.04%".

If, then only "Leave apps in memory" could help you. But the "stuck at 100%" will happen anyway - the problem is checkpointing at the 100% instead of finishing (thus able to immediately report when getting its turn later again).

(Have been testing it extensively with debug version of Rosetta 5.25 on Boinc 5.3.31 and now running on Boinc 5.5.15 - it behaves the same with "Leave apps in memory" option set - no stuck problems except the 100% one.)

Peter

That differs from my experience today, Peter. All my boxes have "Leave in memory" set, including the one that stuck before the end.

Anyway it is good to know that it is already being investigated, and thanks for your response.

If it happens again, do you want further reports or not? Is it useful to know the tasks that fail like this? Are there any files, etc, it would be useful to keep next time?

R~~

I've had quite a few compute errors come through on one of my hosts (haven't checked any of the other hosts yet):

https://boinc.bakerlab.org/rosetta/results.php?hostid=279343

As is Schoasch's computer who's just posted
this thread:

https://boinc.bakerlab.org/rosetta/results.php?hostid=312686

---

(Have been testing it extensively with debug version of Rosetta 5.25 on Boinc 5.3.31 and now running on Boinc 5.5.15 - it behaves the same with "Leave apps in memory" option set - no stuck problems except the 100% one.)

Peter

They're on v5.6.5 now, though development on 5.6 has stopped and they are jumping to 5.7 as of today (to prepare for 5.8 WCG/fancifiedGUI)

---

Team mauisun.org

(Have been testing it extensively with debug version of Rosetta 5.25 on Boinc 5.3.31 and now running on Boinc 5.5.15 - it behaves the same with "Leave apps in memory" option set - no stuck problems except the 100% one.)

That differs from my experience today, Peter. All my boxes have "Leave in memory" set, including the one that stuck before the end.

This might be also because I used (and use) a (75 MB huge :-) debug version from D. Kim - it may actually behave slightly differently from the released one. Previously (with the official one) I observed the same problems as you.

If it happens again, do you want further reports or not? Is it useful to know the tasks that fail like this?

I hope you did not ask me? ;-)
(I was only a yet-another-small-whinner, who helped few devs to make some thoughts and tests.)

Peter

> Over the last few days have seen this error
Incorrect Function exit code 1
Exit at: .initialize.cc line:236

has happened on the following workunits
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029257
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029163
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029157
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029182
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029211
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029218
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029219
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029255
https://boinc.bakerlab.org/rosetta/workunit.php?wuid=36029257

This may help the debugging of 5.32 even though all units are 5.25.

---