reported by Rhiju in Rosetta App Version Release Log

"In Rosetta 5.3.6:

The workunits with varying bond geometry were slow in the previous release, and led to some very uneven crediting. We've put in features that significantly accelerate those workunits. There are also fixes that will allow us to test more accurate energy functions.

There are some fixes for graphics to make the native fold always show up. Finally, we're trying out a new output format that will enable us to send back results for designed proteins (e.g., for HIV vaccines) in compressed format.

The workunits with varying bond geometry were slow in the previous release, and led to some very uneven crediting. We've put in features that significantly accelerate those workunits. There are also fixes that will allow us to test more accurate energy functions.

There are some fixes for graphics to make the native fold always show up. Finally, we're trying out a new output format that will enable us to send back results for designed proteins (e.g., for HIV vaccines) in compressed format."

--------

When posting errors, give links to the machine having the problem. Give links to the WUs that had the errors you're talking about. If your computers are "hidden" consider making them visible while your system is having problems. If you've tried a few diagnostic steps, please mention them. If you're running unusual software or unusual hardware that might possibly cause windows problems then please mention it. For diagnosing.. too much information is much better than too little.

---

Rosetta Moderator: Mod.Tymbrimi
ROSETTA@home FAQ
Moderator Contact

For those of us single (Dual Boot) machines, do we need to give a direct link? My machine shows up as one XP and one Linux machine. Hard to get lost there I hope.

I have them (or it) visible.

---

Keith, if you can provide links to specific WUs, it gives people something specific to discuss. If you just say "I had a failure", we're all left guessing, and perhaps miscommunicating about what may have occured. It takes a little digging on your part to find it... but it saves digging on the part of people that can help you with the problem and everyone that reviews your post.

---

Rosetta Moderator: Mod.Sense

Keith, if you can provide links to specific WUs, it gives people something specific to discuss. ...

Seems to me that the link to the task (result) (not the WU) is what is needed. Usually there is only one task per WU, but after an error there may be more than one. And it is the result not the WU that gives the stderr listing.

Also if you link to the task, both the machine and the workunit can be reached from there in only one more click, so the vital link is the task.

I'd respectfully suggest that if you ask people to post a link to the result in question, they are more likely to get it right than asking for more than one link.

So, for those who do not know how to do this, it is easier to find the right result after the client has reported. One way to do it is as follows (if you prefer another way, keep to it!)

Follow the your acount link from the Rosetta main page. Click on the "View" link for Computers on this Account. Look for the right computer. Click on the results (number forms the link) for that computer. Look for a result with the appropriate run time, error status, etc - probably the last result reported on that computer. Go to that result, copy the address from the address bar, and paste that address into the message here.

Hope that helps
River~~

---

OK. I think this is my first 5.36 error.
https://boinc.bakerlab.org/rosetta/result.php?resultid=45008538

OK I must admit I'm not very well practiced on BOINC forums as far as quoting and inserting links into posts.

Hey, this might be an idea. For people new to BOINC, a link might be provided to learn about BOINC forums.

---

Like this link? BBCode tags. Basically you use BBCode tags to give your post attributes like:

Color
and
Bold text
and
Color and bold
and
links (this one is a junk link

Click "reply to this post" and you can see the tags I used.

basically you type [url={put url addy here}]put name you wish to appear here[/url]

---

I'm not sure how serious this is - but anyway..

one of my hosts was physically removed&replaced and when I checked to see if everything works I pushed the graphics button and then it crashed :

Result.

One other thing that I saw was that FRA_2rio tasks sometime go completley dark for a few minutes in the graphic view fields('Searching_Backbone', 'Accepted' and 'Low Energey' and only 'Native' is shown).
sometimes they only show the edge or another small part of the shape.
than all of a sudden it gets normal again.

I don't know how frequently it happens, but it had happend more than once.

---

Faust.

The seond problem is WU-specific and in the early stage of the simulation, there are some random numbers put in the structure tentatively which will be modeled later. The graphic drawing is currently not smart enough to catch that yet, but I think this will be fixed in the next update.

I'm not sure how serious this is - but anyway..

one of my hosts was physically removed&replaced and when I checked to see if everything works I pushed the graphics button and then it crashed :

Result.

One other thing that I saw was that FRA_2rio tasks sometime go completley dark for a few minutes in the graphic view fields('Searching_Backbone', 'Accepted' and 'Low Energey' and only 'Native' is shown).
sometimes they only show the edge or another small part of the shape.
than all of a sudden it gets normal again.

I don't know how frequently it happens, but it had happend more than once.

---

Native graphic disjointed. There's a section of red hanging off the backbone on my FRA_t362 task.

---

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/

Every one of my results under 5.34 and 5.36 that were successful shows the same stderr out: Meant to say out of the first 20 results.

<core_client_version>5.4.9</core_client_version>
<stderr_txt>
# cpu_run_time_pref: 28800
# random seed: 1776458
# cpu_run_time_pref: 28800
# cpu_run_time_pref: 28800
WARNING! error deleting file .xx2reb.out
======================================================
DONE :: 1 starting structures built 54 (nstruct) times
This process generated 54 decoys from 54 attempts
======================================================


BOINC :: Watchdog shutting down...
BOINC :: BOINC support services shutting down...

</stderr_txt>

Deleting the file from where?

---

Sorry, even if the JPEG link worked without my Yahoo! ID... it isn't clear enough. There is just a hunk of red illustraited completely separately from the rest of the structure in the Native window.

---

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/

Just found one of my machines with a pop-up dialog box (no time to futz with an image, but I can email it if helpful):

Runtime Error!

Program: ...c.bakerlab.orgrosetta_5.36_windows_intelx86.exe

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.

OK

BOINC manager was showing me a single task, no new taskd downloaded, and the CPU was idle.

Messages were:

11/1/2006 8:20:26 PM|rosetta@home|Computation for task FRA_t362_NOCST_hom001_9_t362_6_1s5uA_IGNORE_THE_REST_53_1321_68_0 finished
11/1/2006 8:20:26 PM|rosetta@home|Starting task FRA_2rio_154E_02_hom002_1_2rio_1_1zy4A_IGNORE_THE_REST_127_1327_14_0 using rosetta version 536
11/1/2006 8:20:28 PM|rosetta@home|Started upload of file FRA_t362_NOCST_hom001_9_t362_6_1s5uA_IGNORE_THE_REST_53_1321_68_0_0
11/1/2006 8:20:31 PM|rosetta@home|Finished upload of file FRA_t362_NOCST_hom001_9_t362_6_1s5uA_IGNORE_THE_REST_53_1321_68_0_0
11/1/2006 8:20:31 PM|rosetta@home|Throughput 130050 bytes/sec
11/2/2006 2:20:33 AM|rosetta@home|Sending scheduler request to https://boinc.bakerlab.org/rosetta_cgi/cgi
11/2/2006 2:20:33 AM|rosetta@home|Reason: To report completed tasks
11/2/2006 2:20:33 AM|rosetta@home|Reporting 1 tasks
11/2/2006 2:20:38 AM|rosetta@home|Scheduler request succeeded
11/2/2006 8:39:30 AM|rosetta@home|Unrecoverable error for result FRA_2rio_154E_02_hom002_1_2rio_1_1zy4A_IGNORE_THE_REST_127_1327_14_0 (The system cannot find the path specified. (0x3) - exit code 3 (0x3))
11/2/2006 8:39:30 AM|rosetta@home|Deferring scheduler requests for 1 minutes and 0 seconds
11/2/2006 8:39:30 AM||Rescheduling CPU: application exited
11/2/2006 8:39:30 AM|rosetta@home|Computation for task FRA_2rio_154E_02_hom002_1_2rio_1_1zy4A_IGNORE_THE_REST_127_1327_14_0 finished

When I hit OK, next work unit downloaded and started crunching.

This computer, and I believe this work unit.

BOINC Manager did not shut down, and the machine appears to be crunching. Please let me know if more data is desired or I should take any action.

---

Yes, I'm getting a dismembered native structure too. 3 bits of green disconnected from the main structure.

Task FRA_2rio_154E_02_hom002_1_2rio_1_2bdwA_IGNORE_THE_REST_116_1327_63_0 using rosetta version 536

This one ran more than 30 minutes over my 8 hour preference, which is very unusual. I aborted it.

Here it is.

---

Regards,
Bob P.

Further observations on disjoint structure shown in native box. I'm now observing a FRA_2rio task, and note that the disconnected segments are moving when my "low energy" structure changes. However when I spin the native structure, the detached segments rotate with it. When I spin the low energy structure, they do not move.

Now observing start of a new model on same WU. Seems each of the proteins is off center from the center of my display. Until I got to around step 8,000 the image was often off the page, or only a small portion showing on the side. Rotation of the low energy structure shows it spinning around an axis which is way too large. You have to spin it around off the page to look at the backside. Now, furher in to the model things are back to normal where the spin is centered in the center of the structure, rather then off of it.

---

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/

The file it tries to delete is the output result file from Rosetta. Normally it is gzipped and then the source file is deleted. However, the windows build has a problem of deleting such a file ( or simply issue such a warning) after it is gzipped. We don't see that warning from Linux or Mac build. We will turn it off in the next update as it is harmless ( the output will be returned and validated in a gzip form) but causes confusion.

Every one of my results under 5.34 and 5.36 that were successful shows the same stderr out: Meant to say out of the first 20 results.

<core_client_version>5.4.9</core_client_version>
<stderr_txt>
# cpu_run_time_pref: 28800
# random seed: 1776458
# cpu_run_time_pref: 28800
# cpu_run_time_pref: 28800
WARNING! error deleting file .xx2reb.out
======================================================
DONE :: 1 starting structures built 54 (nstruct) times
This process generated 54 decoys from 54 attempts
======================================================


BOINC :: Watchdog shutting down...
BOINC :: BOINC support services shutting down...

</stderr_txt>

Deleting the file from where?

---

I had also a WU that crashed when using the screen saver. I was busy to rotate the native structure where a red piece was diconnected from the main and

<core_client_version>5.4.11</core_client_version>
<message>
Le chemin d'acc賠sp飩fi頥st introuvable. (0x3) - exit code 3 (0x3)
</message>
<stderr_txt>
# random seed: 1477144
# cpu_run_time_pref: 21600

</stderr_txt>

---

I just had this result hang, and had to terminate it:

result

I walked over to the machine and the screensaver was on, but it wasn't Rosetta, it was Seti. The Seti graphics had frozen, and I couldn't wake up the machine. I hit ctrl-alt-del to get the task manager, and found Rosetta using two cpu's worth of time (out of 4), where normally it would use only 1. The task manager, BTW, was the only thing I could get to come up, and it was sitting on top of the frozen Seti graphics. I killed the Rosetta WU, and immediately the Seti graphics disappeared, and the machine woke up. The Rosetta WU reported "computation error" in the status column.

I suspect this happened when the BOINC manager did its usual 10 minute graphic switchover, but I'm not sure if it was Rosetta trying to take over from Seti, or not letting go when Seti wanted to take over.

---

Yesterday, I got tired of having my computer freezing on the BOINC screensaver with I started using the computer after a break, so I turned off the screensaver. Haven't had a problem since.

Bandit's Mom

---