Ok I finally was able to reload boinc 5.2.8 after trying BOINC 5.2.13. Every R@H Wu is still erroring out so the problem is not with BOINC. The system was running fine until about a week ago. After upgrade to BOINC 5.2.13 R@H began erroring on every WU after about 20 seconds of CPU time. The errors are all the same except for the BOINC version number which of course depends on what is running at the time-

<core_client_version>5.2.8</core_client_version> <message>process exited with code 1 (0x1) </message> <stderr_txt> </stderr_txt>

or

<core_client_version>5.2.13</core_client_version> <message>process exited with code 2 (0x2) </message> <stderr_txt> 2005-12-04 11:34:02 [rosetta@home] execv(../../projects/boinc.bakerlab.org_rosetta/rosetta_4.78_powerpc-apple-darwin) failed: error -1 execv: No such file or directory </stderr_txt>

I have no idea what this directory error means as I have never seen it before. Could someone from the project offer some assistance here?

I have tried posting in the questions and problems/Macintosh area but it seems nobody but the Mac folks ever read or post there.

Regards
Phil

---

We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.

<core_client_version>5.2.13</core_client_version> <message>process exited with code 2 (0x2) </message> <stderr_txt> 2005-12-04 11:34:02 [rosetta@home] execv(../../projects/boinc.bakerlab.org_rosetta/rosetta_4.78_powerpc-apple-darwin) failed: error -1 execv: No such file or directory </stderr_txt>

It looks like something is wrong either with rosetta being able to make the file stderr_tex or to find it. This could be due some faulty assumption we are making about the structure of Mac filesystems vs. Windows.

It looks like something is wrong either with rosetta being able to make the file stderr_tex or to find it.

Snake, have you repaired permissions, verified drive, etc? This sounds a lot like BOINC isn't able to write to the file. I don't think it's a project problem, or we'd (Mac people) all be seeing it...

I'd also advise running Disk Warrior. (NOT, repeat, NOT, TechTool, especially if it's the version on the AppleCare CD, which destroys Tiger disks, unrecoverably!) Your whole difficulty "downgrading" BOINC, plus this, just makes me think there's something wrong on your disk, possibly in the directory structure.

---

It looks like something is wrong either with rosetta being able to make the file stderr_tex or to find it.

Snake, have you repaired permissions, verified drive, etc? This sounds a lot like BOINC isn't able to write to the file. I don't think it's a project problem, or we'd (Mac people) all be seeing it...

I'd also advise running Disk Warrior. (NOT, repeat, NOT, TechTool, especially if it's the version on the AppleCare CD, which destroys Tiger disks, unrecoverably!) Your whole difficulty "downgrading" BOINC, plus this, just makes me think there's something wrong on your disk, possibly in the directory structure.

Well as a matter of fact I also use disk warrior. I usually start from the assumption that the problem is on my end. I have tried everything. I was finally able to downgrade to 5.2.8 and I am still not able to run R@H. It doen not seem to be a permission thing either. If it is the Disk utility can't find the problem.

Regards
Phil

---

We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.

Just a wild question, did you look in the directory to see if the file is there?

---

Just a wild question, did you look in the directory to see if the file is there?

Looking back at the error message, I'm not sure _which_ file is the problem - stderr.txt or the rosetta app itself.

Hm... if the app isn't there, then a "Reset project" would be in order. Failing that, detach and reattach. Leave it to Paul to ask the "is it plugged in" version, when the rest of us are analyzing the power supply... :-)

---

Just a wild question, did you look in the directory to see if the file is there?

Looking back at the error message, I'm not sure _which_ file is the problem - stderr.txt or the rosetta app itself.

Hm... if the app isn't there, then a "Reset project" would be in order. Failing that, detach and reattach. Leave it to Paul to ask the "is it plugged in" version, when the rest of us are analyzing the power supply... :-)

I have reset the project, it did not fix the problem. As to the file name I do not see a file by that name in either the project directory or the boinc data directory. As close as it gets are the files "stderrdae.txt", "stderrgui.txt", Now it would be a surprise to me if the system can't create whatever it needs. But just looking at the error message it would seem it is looking for a directory tree that does not exist. In any case the path in the error message is truncated so I can't even try to create it manually.

All of this is baffling. The software was working just fine at BOINC 5.2.8. I upgraded to the recommended version 5.2.13, and it stopped working. I returned the system to 5.2.8 and it still won't work. This tells me that something got changed by the 5.2.13 install that reverting can't fix by itself. The only question is what.

No if I had to guess it would be that after the upgrade I got a new version of the R@H application from the server (I assume this happens automatically). That new version is somehow not compatible with the G4 dual, but reverting does not replace the 4.78 app with the correct one.

This is really frustrating, because this is one of my fastest machines. While I have not talked to any of the other Mac guys, just on the MACNN team, a number of people are showing "0" credits for the last few days. This is surprising because they were all showing credits in the hundreds just a week ago. So I may not be alone in this.

Regards
Phil

---

We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.

Sounds like you know what you're doing. It certainly sucks to have trouble like this -- especially on your fast box!

It does seem like something changed between 5.2.8 and 5.2.13. When you "reloaded" 5.2.8, did you just install over 5.2.13? I'd try renaming the BOINC folder (to keep everything in case you need it) and install 5.2.13 and/or 5.2.8. See of a totally fresh install will work for either of the versions. If a new install doesn't work, you KNOW it's not an artifact of the upgrade.

---

Just a wild question, did you look in the directory to see if the file is there?

Looking back at the error message, I'm not sure _which_ file is the problem - stderr.txt or the rosetta app itself.

Hm... if the app isn't there, then a "Reset project" would be in order. Failing that, detach and reattach. Leave it to Paul to ask the "is it plugged in" version, when the rest of us are analyzing the power supply... :-)

I have reset the project, it did not fix the problem. < snip >

I found that when I detached and reattached, it redownloaded everything. While I'm not a MAC guru, I gather that MAC OS X is very Unix'esque indeed. Under Unix, the error (execv) means that some program (presumably Boinc) is trying to execute something else (Rosetta), and getting a "No such file .... " error.

That in turn makes me want to agree with Paul that is sounds like the Rosetta app itself has gone AWOL.

Since we're just about to the "try anything once since it can't make it worse" stage, I'm going to suggest a detach and reattach.

---

The file missing is "rosetta_4.78_powerpc-apple-darwin" ... according to the message it should be in the directory: projects/boinc.bakerlab.org_rosetta of the base BOINC directory.

Sorry I was not clearer.

However, a project reset or detach/attach (preferrect) SHOULD have cleared this up.

If it did not, then some more detective work is needed. There are a couple other possibilities. For some of them I would want to rummage through your log files (TXT files in the boinc dir).

The two other possibilities that occur to me are that the client state file is hosed and has bad data. Again, project reset should have cleared that up. For the moment I would discount that.

The other possibility is that there is bad data in the slots directories form past errors.

First, is the file in the rosetta project directory?

I will try to be on today, but, I cannot promise anything ... it is starting to look like a bad day ... maybe I tried to work too much (added to the wiki, beat up Jack about the rnd fuction, etc.).

If you want to try to speed up some things, send me the TXT files zipped up, remind me that this is rosetta problem and which one ... p.d.buck@comcast.net

No promises, but I will try to watch for this ...

---

The file missing is "rosetta_4.78_powerpc-apple-darwin" ... according to the message it should be in the directory: projects/boinc.bakerlab.org_rosetta of the base BOINC directory.

Sorry I was not clearer.

However, a project reset or detach/attach (preferrect) SHOULD have cleared this up.

If it did not, then some more detective work is needed. There are a couple other possibilities. For some of them I would want to rummage through your log files (TXT files in the boinc dir).

The two other possibilities that occur to me are that the client state file is hosed and has bad data. Again, project reset should have cleared that up. For the moment I would discount that.

The other possibility is that there is bad data in the slots directories form past errors.

First, is the file in the rosetta project directory?

I will try to be on today, but, I cannot promise anything ... it is starting to look like a bad day ... maybe I tried to work too much (added to the wiki, beat up Jack about the rnd fuction, etc.).

If you want to try to speed up some things, send me the TXT files zipped up, remind me that this is rosetta problem and which one ... p.d.buck@comcast.net

No promises, but I will try to watch for this ...

I checked this a few times. The Rosetta app is there. Just to be sure, I pulled it from the directory, reset the project and it downloaded a fresh copy. It gave me two new WUs, and sure enough they errored out at 20 seconds CPU time. This has not effected any of the other projects running on this machine. In fact some of them seem to like the newer versions of BOINC.

Based on the error message it seems to be looking for a file that just is not there among the little system files. It seems amazing to me that the system would not create what it needs if something is missing. The only file that seems to cause problems for other apps is the master fetch file, which makes sense because without that BOINC has no idea where to look for stuff.

So Paul, Which .txt files would you want to see. There are a bunch of them in the BOINC Data directory, there are none in the project directory.

Regards
Phil

---

We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.

So Paul, Which .txt files would you want to see. There are a bunch of them in the BOINC Data directory, there are none in the project directory.

Phil,

I use the daemon files stdoutdae.txt, stderrdae.txt

p.d.buck@comcast.net

This is strange ...

---

On Mac's it appears to me not to be the core client but rather the problem is somewhere in the application. I'm a little curious about all the compressed files in>library>application support>Boinc Data>Projects>boinc.bakerlab.org_rosetta> on a G5 using the BM 5.2.13. The same thing also applied when using v4.44.

It appears to me that the Rosetta app works in a totally different fashion than some of the other project apps? I must admit that I'm a little lost.

I am experiencing exactly the same behavior with my PowerMac with dual G5 running at 2 GHz. I have not been able to upload a single result for 4 days. Updating Boinc to 5.2.13 has not had any effect.
If I pause the rosetta@home application there is no problem - seti runs OK.
If i have both rosetta@home and seti@home sharing (50% each) then Boinc locks up and nothing gets crunched.
I have not looked at the log - files yet - will do so later.

---

So Paul, Which .txt files would you want to see. There are a bunch of them in the BOINC Data directory, there are none in the project directory.

Phil,

I use the daemon files stdoutdae.txt, stderrdae.txt

p.d.buck@comcast.net

This is strange ...

Paul,

I get the read you're far more of a Mac guru than I am, so I'll leave it to you to check this out / shoot holes in it / whatever.

One other thing to check. Under Unix (as in Linux / Berkeley) if the execute bit is not set on the file, all bets will be off.

Has anyone checked to see if this is the case? As in, is it showing:

rw-r--r--

or:

rwx-r-xr-x

type of thing. This is Unix-speak, adjust as necessary for the Mac.

---

Assuming that no positive action was taken, the execute bit should not be "damaged". In other words, install BOINC on the Mac, attach to projects and away you go. And I would not call my self a guru ... I leave my computers on all the time as I forget where the power buttons are ...

There *IS* an issue with the current SETI@Home flail-ex ... for some people it is snarfling up the works. As I understand it, JM VII is on the case, but this will not get up where se need to go for this right now.

Some people have had success suspending SETI@Home and then "tapping" their buffer size up, the updating to get more work. But, I am not sure I understand exactly what is being seen here ...

Worse, I am not the best troubleshooter ... Tony is over at SETI@Home ... Though Bill is not bad and he hangs here ...

---

Assuming that no positive action was taken, the execute bit should not be "damaged". In other words, install BOINC on the Mac, attach to projects and away you go. And I would not call my self a guru ... I leave my computers on all the time as I forget where the power buttons are ...

There *IS* an issue with the current SETI@Home flail-ex ... for some people it is snarfling up the works. As I understand it, JM VII is on the case, but this will not get up where se need to go for this right now.

Some people have had success suspending SETI@Home and then "tapping" their buffer size up, the updating to get more work. But, I am not sure I understand exactly what is being seen here ...

Worse, I am not the best troubleshooter ... Tony is over at SETI@Home ... Though Bill is not bad and he hangs here ...

It is definatly not the execute bit. The BOINC downloads the WU, The application starts to crunch (IE it runs), after 20-23 seconds the WU errors out, and it moves to the next WU. This behavior is 100% consistant. The errors would indicat that the application is looking for a file that does not exist, and it apperently can't create the file. There was a simular problem with CP@H and the fix was to build a dummy file tree with an empty file at the bottom for the App to find.

The problem here is that the error message does not give the entire path to the file it is looking for so there is no way to figure out what the system wants or is looking for. The same machine is running S@H, P@H, and CP@H. It is also registered and occasionally runs E@H. None of these other apps are having any problems. BOINC runs just fine as well. There seems to be some problem related to the BOINC 5.2.13 upgrade and the R@H app for the Mac G4 dual.

Regards
Phil

---

We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.

I have detached from Rosetta@Home until something is fixed because I was consistently getting client errors. I'm still running BOINC 4.44 (superbench menubar), so it's not a 5.2.13 issue with my machine.

---

G5 dual 2GHz OSX 10.3.9
G4 dual 867MgHz OSX 10.4.3

It is definatly not the execute bit. The BOINC downloads the WU, The application starts to crunch (IE it runs), after 20-23 seconds the WU errors out, and it moves to the next WU. This behavior is 100% consistant. The errors would indicat that the application is looking for a file that does not exist, and it apperently can't create the file. There was a simular problem with CP@H and the fix was to build a dummy file tree with an empty file at the bottom for the App to find.

The problem here is that the error message does not give the entire path to the file it is looking for so there is no way to figure out what the system wants or is looking for. The same machine is running S@H, P@H, and CP@H. It is also registered and occasionally runs E@H. None of these other apps are having any problems. BOINC runs just fine as well. There seems to be some problem related to the BOINC 5.2.13 upgrade and the R@H app for the Mac G4 dual.

Regards
Phil

I do agree. I have run R@H by it self in BOINC and what is described above happens. However, when running seti@home, BOINC-manager crashes too - some times (3 times in 2 weeks). If I run P@H, S@H and R@H only R@H crashes ...
Now: S@H is having difficulties and I have stopped BOINC altogether to crunch an other project - until this is fixed. Then I will return the PowerMac to R@H ... until then I just use PCs.

---

I have detached from Rosetta@Home until something is fixed because I was consistently getting client errors. I'm still running BOINC 4.44 (superbench menubar), so it's not a 5.2.13 issue with my machine.

If you can still find it try BOINC 5.2.8. There is a known problem running 4.44. super bench with R@H since the last R@H upgrade. R@H only supports 5.2.8 or higher. I had no problems with BOINC 5.2.8 and it actually fixed the problem you are having for my system when I installed it. Then I got over confident in the BOINC folks and installed 5.2.13 which broke the system. Re-installing 5.2.8 did not fix that issue.

Regards
Phil

---

We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.