http://boinc.berkeley.edu/download.php

New Features:

* Grid based views.
* Advanced preferences dialog.
* Project list in the attach to project wizard.
* Improved Windows Vista compatibility.
* An Uninstaller is now included with the Macintosh BOINC Manager distribution.
* Support for x64 on Windows and Linux.

Notes:

· NOTE: BOINC may NOT start at startup if you are running Windows Defender (WinXP, WinXPx64, WinVista, WinVistax64). This issue should be resolved in a future version of BOINC. In order to have BOINC start automatically you'll need to install as a service. On Vista the user account will need to have administrative access to the system.

· See http://boinc.berkeley.edu/linuxx64.php for details on running BOINC on Linuxx64.

Thanks to everyone who tested and logged bugs in this release. We have another good release.

----- Rom

I installed the new client (BOINC 5.10.7) on my Mac. I have 2 things to report.

1) I had to restart the machine as the client generated LOTS of zombie 'ps'
processes that ran the system out of processes.
2) The 'Help' menu in the Boinc manager has a list of 3 items. However,
these items all point to the same place rather than the discreet items
listed.

We know about 2. Can you imagine I have been reporting this since 5.10.0 and they still haven't fixed that? ;-)

---

I installed the new client (BOINC 5.10.7) on my Mac. I have 2 things to report.

1) I had to restart the machine as the client generated LOTS of zombie 'ps'
processes that ran the system out of processes.
2) The 'Help' menu in the Boinc manager has a list of 3 items. However,
these items all point to the same place rather than the discreet items
listed.

This is the first report I've heard of problem (1). A couple of questions about (1):
How quickly did this happen? Was it a meatter of seconds, minutes, hours, or days?
Did the client continue generating zombie processes again after the restart?

(2) Requires a change at the server. I'll remind the appropriate people.

---

Charlie Fenton
BOINC / SETI@home Macintosh & Windows Programmer

(1) Just happened again. A few hours after the restart.

This time I was able to quit BOINC. After a minute or so all the 'ps' processes
cleared out. It had gone several hours with just 2 'ps' processes in the queue
whose PPID was a setiathome process (I have 2 CPUs), then it went bonkers
and there were many 'ps' processes in the queue and the 'no more processes'
error.

Since the processes are zombies they apparently can't be seen via Activity Monitor,
but 'ps -ax' shows them:

14370 ?? RN 0:26.19 setiathome_5.18_powerpc-apple-darwin
14371 ?? RN 0:26.56 setiathome_5.18_powerpc-apple-darwin
14374 ?? ZN 0:00.00 (ps)
14375 ?? ZN 0:00.00 (ps)

Since the processes are zombies they apparently can't be seen via Activity Monitor,
but 'ps -ax' shows them:

14370 ?? RN 0:26.19 setiathome_5.18_powerpc-apple-darwin
14371 ?? RN 0:26.56 setiathome_5.18_powerpc-apple-darwin
14374 ?? ZN 0:00.00 (ps)
14375 ?? ZN 0:00.00 (ps)

The 5.18 version of SETI@home was only recently released to the public (the previous version was 5.13). So this raises the question of whether the problem is BOINC or SETI.

Did you notice if you were successfully running SETI@home 5.18 before upgrading BOINC?
Could you try reverting to an older version of BOINC and see if the problem persists?
What version of BOINC were you running before?

---

Charlie Fenton
BOINC / SETI@home Macintosh & Windows Programmer

Well it has been about a half hour and the zombies are stacking up.

14370 ?? RN 33:05.63 setiathome_5.18_powerpc-apple-darwin
14371 ?? RN 25:12.13 setiathome_5.18_powerpc-apple-darwin
14374 ?? ZN 0:00.00 (ps)
14375 ?? ZN 0:00.00 (ps)
14609 ?? ZN 0:00.00 (ps)
14610 ?? ZN 0:00.00 (ps)
14616 ?? ZN 0:00.00 (ps)
14617 ?? ZN 0:00.00 (ps)
14621 ?? ZN 0:00.00 (ps)
14622 ?? ZN 0:00.00 (ps)
14626 ?? ZN 0:00.00 (ps)
14627 ?? ZN 0:00.00 (ps)
14631 ?? ZN 0:00.00 (ps)
14632 ?? ZN 0:00.00 (ps)
14636 ?? ZN 0:00.00 (ps)
14637 ?? ZN 0:00.00 (ps)

Poking around I found these lines in the stderrae.txt file

-s: illegal process id: KILL
-s: illegal process id: KILL
-s: illegal process id: KILL
-s: illegal process id: KILL
-s: illegal process id: KILL
-s: illegal process id: KILL
-s: illegal process id: KILL
-s: illegal process id: KILL
-s: illegal process id: KILL

Since the processes are zombies they apparently can't be seen via Activity Monitor,
but 'ps -ax' shows them:

14370 ?? RN 0:26.19 setiathome_5.18_powerpc-apple-darwin
14371 ?? RN 0:26.56 setiathome_5.18_powerpc-apple-darwin
14374 ?? ZN 0:00.00 (ps)
14375 ?? ZN 0:00.00 (ps)

The 5.18 version of SETI@home was only recently released to the public (the previous version was 5.13). So this raises the question of whether the problem is BOINC or SETI.

Did you notice if you were successfully running SETI@home 5.18 before upgrading BOINC?
Could you try reverting to an older version of BOINC and see if the problem persists?
What version of BOINC were you running before?

I was running BOINC 5.8.17. Not sure aboiut SETI@home, but I do think it was 5.13.
Are they any issues going back to BOINC 5.8.17?
I am going to have to do something in order to keep the machine running overnight.

Are they any issues going back to BOINC 5.8.17?

On Windows machines you have to uninstall the newer version and reinstall the older version. Seeing that the Mac version now has its own installer, I assume this would work the same.

---

Since the processes are zombies they apparently can't be seen via Activity Monitor,
but 'ps -ax' shows them:

14370 ?? RN 0:26.19 setiathome_5.18_powerpc-apple-darwin
14371 ?? RN 0:26.56 setiathome_5.18_powerpc-apple-darwin
14374 ?? ZN 0:00.00 (ps)
14375 ?? ZN 0:00.00 (ps)

The 5.18 version of SETI@home was only recently released to the public (the previous version was 5.13). So this raises the question of whether the problem is BOINC or SETI.

Did you notice if you were successfully running SETI@home 5.18 before upgrading BOINC?
Could you try reverting to an older version of BOINC and see if the problem persists?
What version of BOINC were you running before?

I was running BOINC 5.8.17. Not sure aboiut SETI@home, but I do think it was 5.13.
Are they any issues going back to BOINC 5.8.17?
I am going to have to do something in order to keep the machine running overnight.

Well that was a waste of time and effort.

I downloaded BOINC 5.8.17 and re-installed. I got the expected versioin of BOINC but I still
have SETI@home 5.18. It did not go back to the older version of SETI. So now I am running
the old BOINC with the apparently misbehaving SETI@home. I still see the zombie 'ps' processes.

Restarting BOINC removes the zombie 'ps' processes from the process
queue, but apparently doesn't free up the memory. I had to restart to
recover the memory.

Anyone know how to get the old SETI@home code? As I said earlier I
get SETI 5.18 with the old version of BOINC.

I downloaded BOINC 5.8.17 and re-installed. I got the expected versioin of BOINC but I still
have SETI@home 5.18. It did not go back to the older version of SETI. So now I am running
the old BOINC with the apparently misbehaving SETI@home. I still see the zombie 'ps' processes.

Ho, there is a big difference.
BOINC is one thing, any project is another.

Do not try to compare the BOINC client number with that of any project's science application number. They may be the same at times, but they may also differ hugely.

Seti 5.18 is the latest Seti application number. It's got nothing to do with the BOINC client number. You won't find BOINC 4.45 download only Seti 4.xx applications. They will download 5.18 as well.

And if this is the last Seti application ever to be released, then even in BOINC 6.2.5 you will still be running Seti 5.18

Just to give an example.

BOINC is always evolving. The new version of BOINC is 5.10.7, the next new version will be a 6.0.x version. Yet the version number of BOINC has nothing to do with the version number of the science application that the project is using.

Some projects are now only starting up and using a 0.9 version. Don't expect them to jump to a 5.xx version of their application overnight. Hopefully it'll take them a couple of years. :)

---

I downloaded BOINC 5.8.17 and re-installed. I got the expected versioin of BOINC but I still
have SETI@home 5.18. It did not go back to the older version of SETI. So now I am running
the old BOINC with the apparently misbehaving SETI@home. I still see the zombie 'ps' processes.

Ho, there is a big difference.
BOINC is one thing, any project is another.

Do not try to compare the BOINC client number with that of any project's science application number. They may be the same at times, but they may also differ hugely.

Seti 5.18 is the latest Seti application number. It's got nothing to do with the BOINC client number. You won't find BOINC 4.45 download only Seti 4.xx applications. They will download 5.18 as well.

And if this is the last Seti application ever to be released, then even in BOINC 6.2.5 you will still be running Seti 5.18

Just to give an example.

BOINC is always evolving. The new version of BOINC is 5.10.7, the next new version will be a 6.0.x version. Yet the version number of BOINC has nothing to do with the version number of the science application that the project is using.

Some projects are now only starting up and using a 0.9 version. Don't expect them to jump to a 5.xx version of their application overnight. Hopefully it'll take them a couple of years. :)

I fully understand this. My point is that going to SETI site to get a download simply directs you
to the BOINC page. So, I assumed if I downloaded the version of BOINC I was running this
morning I would also get the version of SETI I was running this morning.
I got this new version of SETI by downloading the new version of BOINC, so I assumed I could
back to the state of things as of this morning by downloading the previous version of BOINC.
Apparently it is not easy to go back with versions of SETI.

I fully understand this.

I'm sorry if my explanation wasn't complete, but as it would seem you don't understand. And So I will try again.

The application that Seti brings out is independent from the version of BOINC.
Seti releases its own versions of its application, without looking at the BOINC version.

The other way around is true as well. BOINC releases new versions of itself without checking what version a certain application in any of the 51+ BOINC projects may be at.

You won't be able to tell the project what application you want to download. When a project releases a new application, you get it downloaded through BOINC automatically.

---

I fully understand this.

I'm sorry if my explanation wasn't complete, but as it would seem you don't understand. And So I will try again.

The application that Seti brings out is independent from the version of BOINC.
Seti releases its own versions of its application, without looking at the BOINC version.

The other way around is true as well. BOINC releases new versions of itself without checking what version a certain application in any of the 51+ BOINC projects may be at.

You won't be able to tell the project what application you want to download. When a project releases a new application, you get it downloaded through BOINC automatically.

Yes I did understand. However, the current situation seems to require a rollback to SETI 5.13
and the SETI download (as you point out) just points to BOINC download. There does not seem
to be a way to get back to SETI 5.13 although it is trivial to switch BOINC versions without problem.

Since this machine has a problem using SETI 5.18 and there doesn't seem to be a way to get back
to SETI 5.13 I am going to have to quit BOINC so that the machine runs through the night.

However, the current situation seems to require a rollback to SETI 5.13
and the SETI download (as you point out) just points to BOINC download. There does not seem to be a way to get back to SETI 5.13 although it is trivial to switch BOINC versions without problem.

No, Charlie (the developer) asked you to revert to the old version of BOINC that you used prior to this:

Could you try reverting to an older version of BOINC and see if the problem persists? What version of BOINC were you running before? (Post in this thread.)

Since this machine has a problem using SETI 5.18 and there doesn't seem to be a way to get back to SETI 5.13 I am going to have to quit BOINC so that the machine runs through the night.

Now that's a weird suggestion. BOINC <> Seti. In other words, Seti isn;t the only project you can run through BOINC, there are at last count 51 prjects available through BOINC. Ok, Not all with Mac option, but about 33% of them do allow Macs.

---

However, the current situation seems to require a rollback to SETI 5.13
and the SETI download (as you point out) just points to BOINC download. There does not seem to be a way to get back to SETI 5.13 although it is trivial to switch BOINC versions without problem.

No, Charlie (the developer) asked you to revert to the old version of BOINC that you used prior to this:

Could you try reverting to an older version of BOINC and see if the problem persists? What version of BOINC were you running before? (Post in this thread.)

Since this machine has a problem using SETI 5.18 and there doesn't seem to be a way to get back to SETI 5.13 I am going to have to quit BOINC so that the machine runs through the night.

Now that's a weird suggestion. BOINC <> Seti. In other words, Seti isn;t the only project you can run through BOINC, there are at last count 51 prjects available through BOINC. Ok, Not all with Mac option, but about 33% of them do allow Macs.

This where the discussion became unclear.

I downloaded and re-installed BOINC 5.8.17. I was expecting that I would also get back to SETI 5.13
but apparently there is some other (than user download) mechanism in place to update the SETI code.
In any event, I was able to run BOINC 5.8.17 with SETI 5.18, however, after 2 hours of running I get
these zombie 'ps' processes.

UID PID PPID CPU PRI NI VSZ RSS WCHAN STAT TT TIME COMMAND
31 356 348 0 12 19 235812 66128 - RN ?? 83:18.32 setiath
31 357 348 0 12 19 106632 67536 - RN ?? 80:20.12 setiath
0 3839 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3840 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3844 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3845 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3850 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3851 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3855 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3856 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3860 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3861 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3913 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)
0 3914 356 0 0 19 0 0 - ZN ?? 0:00.00 (ps)

Also, somewhere, there is memory that is not being released.

I'm running OSx 10.4.10 on a 2 cpu G5 with 2GB of memory.

I downloaded and re-installed BOINC 5.8.17. I was expecting that I would also get back to SETI 5.13

I'm sorry if I confused you. There is absolutely no way for you to revert to the older version of SETI. Since, by pure coincidence, you started running a new version of SETI@home and also a new version of BOINC, I asked you to revert BOINC so we could see if the problem is with BOINC or with SETI.

Well that was a waste of time and effort.

Acually, it was quite worthwhile. My purpose in asking you to revert was not that I expected it to fix your problem, but to help us analyze what is wrong. And that test showed the problme is not caused by the new version of BOINC.

As you discerned, the fact that the problem continues to happen with the older version makes us suspect the new version of SETI@home.

My own dual G4 has just started running its first SETI@home 5.18 work unit, so I will watch it and see if I observe the problem here.

You might consider attaching to a different project (there are many to choose from) and temporarily suspending SETI@home in the BOINC Manager's Projects tab. Please let me know if you see the problem with other projects.

Also, you might try using "ps -acjx" which will identify the user and the parent . If the zombie application was SETI@home, the user should be boinc_project.

I see you did run ps with the -alx option. The zombies all have a user ID of 0, which is root. I'm not certain, but I believe they would have a user ID of 31 if they were remnants of SETI@home. So it may be that the zombies are not being caused by BOINC or SETI@home. If you quit BOINC for the night, please check whether you still get zombies.

Please let us know what you find out. Thanks for your help.

Cheers,
--Charlie

---

Charlie Fenton
BOINC / SETI@home Macintosh & Windows Programmer