[Narwhal]

Just me again, you know, the guy from Team Picard. We'd like you to come on over to our web site and pay us a visit. www.teampicard.net Hope to see you there.

[Lateralus]

[reader50]

I already paid, and you've already seen me there.

[ImJake]

Nice place you have here! I am Jake, another fellow Picardian. Just thought I would stop by say hello, read up on what you guys are doing.

[Thyme Lawn]

Thyme Lawn checking in to the Picard away team.

Arrived late in the transporter room and they kept me waiting. Something about a load of Tribbles they had to get off the Enterprise before they started breeding too quickly

Nice place you've got here and I love your stats. Hope you don't mind if I wander around a bit ...

... and I promise that no Tribbles sneaked over with me

[Shaktai]

Welcome, welcome all, to the home of the best Mac team ever.

Have a look around and be comfortable. In fact feel free to bring you computers along and stay awhile.

[[APi]TheMan]

You guys are very friendly, I usually end up following one of your sigs over to your place and looking around for awhile... Heh, I hope you guys do the same.

[Thyme Lawn]

Originally posted by [APi]TheMan:
You guys are very friendly, I usually end up following one of your sigs over to your place and looking around for awhile... Heh, I hope you guys do the same.

Hint taken

Adium X looks good ...

... but unfortunately I'm one of those unlucky Mac-less souls

[ditchhook]

I have available to me 10 G-3 I-books running 10.2, and 5 G-4 I-books running 10.4. 85% of the time they sit in a locked cabinet, ten percent of the time they sit idle on a table. Five percent (at most) of the time they are used.

I struggled quite a bit with installing BOINC on these machines. I got my prototype to the point of downloading work units, but never could get it crunching, much less routinely doing so without constant supervision. I've decided I can't get this done alone.

Getting to the point of this post: If someone would be able to provide me with the aid necessary to get these machines crunching, I'd be willing to join whatever team this someone asked me to join. And that team would would get as a bonus the credits for the three PC machines I already have on-line with BOINC.

Whether it be team Picard or MACNN makes no difference to me.

[reader50]

Originally posted by ditchhook:
I have available to me 10 G3 iBooks running 10.2, and 5 G4 iBooks running 10.4.
...

Man, am I getting behind the curve with 10.2.8. Congrats on getting copies of Tiger so early.

Nice post, btw. Once our BOINC experts get home from work, expect a ton of help and advice from members of both teams.

In the meantime, we have several BOINC threads going here. I know that install instructions were posted in more than one of them.

[mikkyo]

Originally posted by ditchhook:
I have available to me 10 G-3 I-books running 10.2, and 5 G-4 I-books running 10.4.

The G3 i-books are much better off with 10.3.x, I would upgrade them if you can.
You could try the 10.2.8 client on them, but folks have varying results.

You can run the Tiger preview client on the 10.4 machines.

Setup is covered in several other threads, but isn't that hard.
Your have an easier time since you can setup one machine, and then copy the Boinc directory to all the other machines.

See this thread for info (note versions have updated since it was written to boinc 4.06.
http://forums.macnn.com/showthread.p...hreadid=219776

BTW, if these machines are only on 10% of the time, you won't get any credit most likely.
All the Boinc projects have WU deadlines, and older machines(G3), when crunching full time, can make the deadlines without problem, 80% of the time would be close.
10% is just not gonna happen. Seti gives you like 2 weeks, I think, before it re-releases the work. Climate Prediction is too intensive for them. Predictor might work, depending on their time requirements.
This sounds like a school setup, and if so, you have 5 days a week for maybe 6 hours max a day. Probably not worth the setup trouble to find out.

If you could keep the units running all night or something, then you have a chance.
I would try installing the 10.4 preview client(fastest one of the bunch) on one of the Tiger machines and setting it up to test it out. See how much crunch time to you get and if credit ever shows up.
Then you can better decide if the others are worth setting up.

[Thyme Lawn]

Congratulations on beating us to answering ditchhook's query.

I did think about pointing him/her towards some of your team's answers on the CPDN BOINC boards, but didn't think that would have been very fair of me

[ditchhook]

Hi again,

First, as mistake I on my part needs to be corrected earlier I claimed to have available five I-Book G4s running OS 10.4. But actually they are 10.3.4.

Second, I can't seem to get the MacNN Boinc 4.05versions available here to open and unpact via stuffit: when I click on the links I get text files.

Third: I went to climateprediction.net to download their MAC X version. (It's the only one up at the moment.) It "runs" on my test machine but never can do a work unit. Here is one cycle of the detail response:

2004-09-04 19:09:23 [climateprediction.net] Starting result 1p7z_200100156_0 using hadsm3 version 4.03
Starting model in /Applications/boinc/projects/climateprediction.net...
Archive: hadsm3data_4.03_powerpc-apple-darwin.zip
creating: 1p7z_200100156/datain/
creating: 1p7z_200100156/datain/ancil/
creating: 1p7z_200100156/datain/ancil/ctldata/
inflating: 1p7z_200100156/datain/ancil/ctldata/spec3a_lw_3_asol2c_hadcm3
inflating: 1p7z_200100156/datain/ancil/ctldata/spec3a_sw_3_asol2b_hadcm3
creating: 1p7z_200100156/datain/ancil/ctldata/stasets/
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01001218
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01002207
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003236
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003237
extracting: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003254
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003255
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003274
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003275
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003276
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003277
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003278
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003279
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003280
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003281
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01003286
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01005207
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01005208
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01005222
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01005223
inflating: 1p7z_200100156/datain/ancil/ctldata/stasets/X01010206
creating: 1p7z_200100156/datain/ancil/ctldata/STASHmaster/
inflating: 1p7z_200100156/datain/ancil/ctldata/STASHmaster/STASHmaster_A
inflating: 1p7z_200100156/datain/ancil/ctldata/STASHmaster/STASHmaster_O
inflating: 1p7z_200100156/datain/ancil/ctldata/STASHmaster/STASHmaster_S
inflating: 1p7z_200100156/datain/ancil/ctldata/STASHmaster/STASHmaster_W
inflating: 1p7z_200100156/datain/ancil/qrclim.icedp.32
inflating: 1p7z_200100156/datain/ancil/qrclim.newsst5.32
inflating: 1p7z_200100156/datain/ancil/qrclim.ozone_preind_corr
inflating: 1p7z_200100156/datain/ancil/qrclim.uvcurr.32
creating: 1p7z_200100156/datain/dumps/
inflating: 1p7z_200100156/datain/dumps/slab32_1810.start
inflating: 1p7z_200100156/datain/lats
inflating: 1p7z_200100156/datain/ppcodes
creating: 1p7z_200100156/dataout/
extracting: 1p7z_200100156/dataout/thist
creating: 1p7z_200100156/jobs/
inflating: 1p7z_200100156/jobs/control.stashc
inflating: 1p7z_200100156/jobs/double.stashc
inflating: 1p7z_200100156/jobs/Recona.12
inflating: 1p7z_200100156/jobs/Recona.13
inflating: 1p7z_200100156/jobs/spec3a_lw_3_asol2c_hadcm3
inflating: 1p7z_200100156/jobs/spec3a_sw_3_asol2b_hadcm3
inflating: 1p7z_200100156/jobs/spin.stashc
inflating: 1p7z_200100156/jobs/yabsd.ihist
inflating: 1p7z_200100156/jobs/yabsd.PRESM_A
extracting: 1p7z_200100156/jobs/yabsd.PRESM_O
extracting: 1p7z_200100156/jobs/yabsd.PRESM_S
extracting: 1p7z_200100156/jobs/yabsd.PRESM_W
inflating: 1p7z_200100156/jobs/yabsd.stashc
inflating: 1p7z_200100156/registration_license.txt
extracting: 1p7z_200100156/stderr_um.txt
extracting: 1p7z_200100156/stdout_um.txt
creating: 1p7z_200100156/tmp/
extracting: 1p7z_200100156/tmp/pipe_dummy
creating: 1p7z_200100156/viz/
inflating: 1p7z_200100156/viz/globe.rgb
Archive: 1p7z_200100156.zip
inflating: 1p7z_200100156/jobs/climate.spin
inflating: 1p7z_200100156/jobs/climate.cont
inflating: 1p7z_200100156/jobs/climate.doub
inflating: 1p7z_200100156/jobs/ncatts.cpdc
inflating: 1p7z_200100156/jobs/spec3a_sw_3_asol2b_hadcm3
Created shared memory region key = 25115
Env Used=DYLD_LIBRARY_PATH=/Applications/boinc/projects/climateprediction.net:../
Copying files for startup...
In pre_initialise_phase (part 1 of 3)
In initialise_phase (part 2 of 3)
In startup_phase (part 3 of 3)
Starting model ID 1p7z_200100156 Phase 1
Stack size=48.00 MB
Waiting for model startup, this may take a minute...
1p7z_200100156 - PH 1 TS 000001 - 00/00/0000 00:00 - H:M:S=0000:00:00 AVG= 0.00 DLT= 0.00
Model crashed...retrying...restart level 0
Preparing for restart...
Rewinding a model-day...
Starting model ID 1p7z_200100156 Phase 1
Stack size=48.00 MB
Waiting for model startup, this may take a minute...
1p7z_200100156 - PH 1 TS 000001 - 00/00/0000 00:00 - H:M:S=0000:00:00 AVG= 0.00 DLT= 0.00
Model crashed...retrying...restart level 1
Preparing for restart...
Rewinding a model-month...
Error: Restart files for dataout/restart.month not found
Giving up, this result exceeded crash count for available restart files.
adding: ncatts.cpdc (deflated 72%)
adding: climate.cont (deflated 79%)
adding: climate.cpdc (deflated 79%)
adding: climate.doub (deflated 78%)
adding: climate.spin (deflated 79%)
adding: 1p7z_200100156.xml (deflated 65%)
adding: ncatts.cpdc (deflated 72%)
adding: ncatts.cpdc (deflated 72%)
adding: ncatts.cpdc (deflated 72%)
adding: stderr_um.txt (stored 0%)
adding: yabsd.out (deflated 93%)
adding: restart.day (deflated 43%)
2004-09-04 19:09:35 [climateprediction.net] Unrecoverable error for result 1p7z_200100156_0 (process exited with code 251 (0xfb))
2004-09-04 19:09:35 [climateprediction.net] Unrecoverable error for result 1p7z_200100156_0 (process exited with code 251 (0xfb))
2004-09-04 19:09:35 [climateprediction.net] Deferring communication with project for 1 minutes and 0 seconds
2004-09-04 19:09:35 [climateprediction.net] Deferring communication with project for 1 minutes and 0 seconds
2004-09-04 19:09:35 [climateprediction.net] Computation for result 1p7z_200100156 finished
2004-09-04 19:09:36 [climateprediction.net] Started upload of 1p7z_200100156_0_1.zip
2004-09-04 19:09:36 [climateprediction.net] Started upload of 1p7z_200100156_0_2.zip
2004-09-04 19:09:36 [climateprediction.net] Finished upload of 1p7z_200100156_0_1.zip
2004-09-04 19:09:36 [climateprediction.net] Throughput 3956 bytes/sec
2004-09-04 19:09:36 [climateprediction.net] Started upload of 1p7z_200100156_0_3.zip
2004-09-04 19:09:37 [climateprediction.net] Finished upload of 1p7z_200100156_0_2.zip
2004-09-04 19:09:37 [climateprediction.net] Throughput 109666 bytes/sec
2004-09-04 19:09:37 [climateprediction.net] Started upload of 1p7z_200100156_0_4.zip
2004-09-04 19:09:37 [climateprediction.net] Finished upload of 1p7z_200100156_0_3.zip
2004-09-04 19:09:37 [climateprediction.net] Throughput 4110 bytes/sec
2004-09-04 19:09:37 [climateprediction.net] Started upload of 1p7z_200100156_0_5.zip
2004-09-04 19:09:38 [climateprediction.net] Finished upload of 1p7z_200100156_0_4.zip
2004-09-04 19:09:38 [climateprediction.net] Throughput 4099 bytes/sec
2004-09-04 19:09:38 [climateprediction.net] Finished upload of 1p7z_200100156_0_5.zip
2004-09-04 19:09:38 [climateprediction.net] Throughput 130055 bytes/sec
2004-09-04 19:10:36 [---] Insufficient work; requesting more
2004-09-04 19:10:36 [climateprediction.net] Requesting 1728000 seconds of work
2004-09-04 19:10:36 [climateprediction.net] Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
2004-09-04 19:10:37 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded
2004-09-04 19:10:37 [climateprediction.net] Message from server: No work available (daily quota exceeded)
2004-09-04 19:10:37 [climateprediction.net] Message from server: No work available (daily quota exceeded)
2004-09-04 19:10:37 [climateprediction.net] No work from project
2004-09-04 19:10:37 [climateprediction.net] No work from project
2004-09-04 19:10:37 [climateprediction.net] Deferring communication with project for 1 hours, 0 minutes, and 0 seconds



Any help? As I promised before, I join the team of the person(s) that help the most to get this going.

[Shaktai]

Okay, for mikkyo's compiles, I have the same problem you do in downloading them.

The fix is to Control-Click on the link, and then choose "Download linked file as..." This will let you save it to your hard-drive in a way that stuffit can recognize. If stuffit has problem with the file, you may need to update your copy of stuffit expander. (Free).

Now as to problem you are having, it has been happening to a lot of people do to some work unit errors. Several suggestions have been made. What worked for me on one of my Windows machines is:
1.) Download the client from the SETI site instead, or now that you know how to download mikkyo's compiles just do that.
2.) Reinstall the client and then detach-reattach to the project. In this case, since you don't have any working units, you can just trash what you have, and start clean. It won't let you get any more units today on the current node, because you have already exceeded the quota due to the number of errors.

Climate Prediction hadn't really been "volume" tested yet. But with SETI's issues and the temporary shutdown of Predictor while they convert over and get their new hardware online, Climate Prediction got slammed with new users. I think that more then just a few work units got corrupted. They are working it out, and the above steps should let you get going.

Hang in there!

[reader50]

In most browsers, option-click on a link to download it to your default download location.

[ditchhook]

The compiled BOINC options are CPU dependent, but I don't know how to query my machine to find out what CPU it holds. "System Profiler" says "PowerPC G4 (1.1)."

[mikkyo]

Some of my machines are finally picking up working models for climate prediction now.
You may have to do the detach/attach thing.
The easiest way I have found to do this is to change the file name of the account_climateprediction.net.xml to something else like "cp".
Then detach the project
./boinc_4.06_ppcGX -detach_project http://climateprediction.net
then just move the "cp" file back to the right name, and start the boinc client anew.
This way you don't have to type in your accountid, as it is stored in the account file.

[Shaktai]

Originally posted by ditchhook:
The compiled BOINC options are CPU dependent, but I don't know how to query my machine to find out what CPU it holds. "System Profiler" says "PowerPC G4 (1.1)."

This link might help you to identify it by speed, and year. The G4's are listed down towards the bottom.
http://www.netneurotic.de/mac/powerpc.html

[ditchhook]

Thanks, because my machine rates itself at 1.07 GHz, and was purchased in the 2004, I'm going to assume it is a 7450 CPU. I installed this version of BOINC, but I have again exceeded my daily quota of requests for wus from cp, so I guess I will see if it works any better tomorrow.

[mikkyo]

Just type in Terminal
hostinfo
and get the info

Mach kernel version:
Darwin Kernel Version 7.5.0:
Thu Aug 5 19:26:16 PDT 2004; root:xnu/xnu-517.7.21.obj~3/RELEASE_PPC


Kernel configured for a single processor only.
1 processor is physically available.
Processor type: ppc750 (PowerPC 750)
Processor active: 0
Primary memory available: 384.00 megabytes.
Default processor set: 54 tasks, 125 threads, 1 processors
Load average: 1.48, Mach factor: 0.07



If you detach and then attach to climateprediction, you can try again.

[ditchhook]

I went through half a dozen or more tries, and was about to give up on getting a climate wu working on my prototype I-book (even so far as to begin this message) when it seems I downloaded a wu that didn't crash: 33 time steps so far!

I'm going to let this run overnight, then try to create and install the auto-restart "crontab" process as described on another thread.

Thanks for the reassurance. After six or so failures, I would have assumed it was my problem and not climate-prediction's!

[Shaktai]

Originally posted by ditchhook:
I went through half a dozen or more tries, and was about to give up on getting a climate wu working on my prototype I-book (even so far as to begin this message) when it seems I downloaded a wu that didn't crash: 33 time steps so far!

I'm going to let this run overnight, then try to create and install the auto-restart "crontab" process as described on another thread.

Thanks for the reassurance. After six or so failures, I would have assumed it was my problem and not climate-prediction's!

Nope, not your problem. They had some very real work unit issues. Glad you are up and running. It will take a while to crunch that unit, so you are good for a time. However every 10,000 or so timesteps (can't remember the exact number) it will trickle credit for the portion of the work unit completed.

[mikkyo]

It is a close competition between seti and climate prediction for the most problematic boinc based DC project.
I have tried to get several machines to start a model working with climate prediction and failed more times than I can count,
I merged about 15 hosts into one on my computer page there just today, due to all the repeated detach/attach attempts.
The success rate is about as random as you describe.
My iBook miraculously picked up a working model after several days of failures.
That doesn't bode well for when it finishes the model, as it might not get another working one for who knows how long. Good thing the models take so long.

The problems appear to be within the models themselves, and they admitted having some bad file in some recent models, but that was supposedly fixed.
Unfortunately, I get the feeling very few folks over there are involved with the technical side of the project as questions go unanswered for days, and it seem about 2 people are on the message board that are directly involved with the project.

There is no list of the model error codes available, yet hundreds of machines are getting them.
Hopefully once the weekend is over there will be some useful activity over there and maybe another nice "ooops" message will go out about some other kind of corrupted model problem, although that doesn't instill any faith in the future of the project.

The scary part is their Windows workers seem to have much more success than their unix(and macosx) workers do.

I can't wait for Predictor to start up again, they have been very proactive with regard to any problems that have come up and keeping clients crunching.

[Shaktai]

Originally posted by mikkyo:
I can't wait for Predictor to start up again, they have been very proactive with regard to any problems that have come up and keeping clients crunching.

I agree there completely. The Predictor team has been much more proactive and communicative then most. If they get their new hardware in on time, wouldn't surprise me to see them back up and running by the end of the week.

FYI: The windows boxes aren't having much more success then the Mac & Linux boxes. Took me 3-4 days just to get a usable unit for my Athlon 1900. My Athlon 2500 crashed or froze so many times, that it wiped out the OS, and was unrepairable. (Never happened on Predictor) Climate Prediction appeared stable until Predictor & SETI went down for upgrades. Then they got a huge influx of people, and apparently their system is struggling with it.

Going to see if I can get Linux running on the 2500 instead of Windows. Getting sick and tired of having to call Microsoft everytime I have to re-install the OS. What would you recommend? Needs to be user friendly, you know I am not all that technical.

[reader50]

You could try installing Darwin x86 if the client binaries are compatible. A BSD x86 binary might work just fine on Darwin x86.

I've long been curious about x86 Darwin, it should behave much like the OSX Terminal, without the higher GUI functions or the Finder services.

[Shaktai]

Originally posted by reader50:
You could try installing Darwin x86 if the client binaries are compatible. A BSD x86 binary might work just fine on Darwin x86.

I think that might be a little beyond me..... I have decided to download a few of the Linux distro iso images that are configured to run from a CDROM. This way I can try them, and figure out which I am comfortable with. SUSE seems to be having trouble with my hardware, probably the mobo, but I did get an older version of Mandrake to run (though with a few difficulties), so am downloading their CD iso to give the latest version it a try. (it's at least 2 full versions newer then what I have now)

[reader50]

Here is what looks like a compressed ISO image. Darwin 7.0.1 (OSX 10.3) for PPC and x86. It must be a dual boot CD.

http://www.opendarwin.org/downloads/...win-701.iso.gz

Release notes (for Darwin 6.x) are here.

[mikkyo]

I put Mandrake 10 on my linux boxes recently.
It has a pretty user-friendly OS Install and decent help and config tools.
None of my linux boxes have picked up a working climate prediction model though.
Darwin is very useable too.
Basically Mac OS X without a GUI.
I have no idea how well it performs when compared to other nixes though.
I used to be a Redhat person, but migrated once they became too commercial.

Maybe Ill make some super optimized i686 compiles of Boinc...

[Shaktai]

While not immediatly clear, there are actually two different ISO images for download. one for PPC and one for X86. Looks like an early version, circa 2002. The read me states specifically that AMD (athlon) is not supported.

Got Mandrake CD-Rom version working fine. If it is stable and able to get some work (got a no work available message) then I'll get the real version and install on the hard-drive. It has an almost friendly GUI, which is what this well worn brain needs.

[mikkyo]

Looks like 2 of my linux boxes got working models, finally.
The macs are still getting the full gambit of errors though.

[ditchhook]

I'm trying to be stoic about the problems BOINC projects are having. If one wants a relatively stable, reliable distributed computing project, seti-classic is out there, among others.

But someone has to help work the kinks out of new projects. Many times problems of scale don't show themselves unless and until one scales up the project.

On the other hand, I do sometimes feel we don't quite get the respect and acknowledgement appropriate to our contribution. Lately I maybe put a couple (donated) hours a week into installing, reinstalling, and monitoring BOINC on my machines. Some project managers don't seem to recognize this contribution of free (and often highly skilled) worker-hours to their projects. This contribution-- in the short run-- is much more significant than the idle cpu cycles. All the work you MAC people are putting into BOINC doesn't appear to be adequately recognized, in my opinion.

I don't appreciate it when project managers don't immediately let me know-- even via their web pages-- what is up, why things aren't working, when they might be working again, and what they need me to do to help them, so that I can figure out how best to use my time. The best non-profits respect their volunteer workers better than this. They wouldn't have volunteers if they didn't. And I suspect Seti@home is losing volunteers at a wholesale rate to other projects.

I would suggest they hire (even part time, or shared amongst the projects) a volunteer coordinator/advisor from the non-profit sector to advise them and be their liason with the volunteers: keeping everyone informed, etc.

[ditchhook]

I noticed that when I simply close my I-Book (lock the screen down over the keyboard) that processing stops. Is there anyway I can keep BOINC going even when the machine is "closed" but still on?

[Nacente]

Don't do it, iBook cools by the keyboard. If you do that for a long time, it'll burn.

[ditchhook]

<B>Don't do it, iBook cools by the keyboard. If you do that for a long time, it'll burn.</B>

Thanks, I didn't realize that.

Another question:

I found on another thread instructions on how to auto restart BOINC:

<b>Make a plain text file with

@reboot cd /Applications/Boinc; /usr/bin/nohup ./boinc_command &

Where the /Applications/Boinc is the path to your boinc folder and ./boinc_command is the boinc command you run (./boinc_3.20_powerpc-apple-darwin or the like).
Save the file in plain text as boinctab.
Then do
crontab boinctab
in Terminal.</b>

While testing BOINC/climateprediction since yesterday, I noticed it crashed on me once in a 24 hour period, so I would like to create a "boinctab" routine that also restarts BOINC once an hour. Anyone done something like that in the past?

[Shaktai]

Originally posted by ditchhook:
I'm trying to be stoic about the problems BOINC projects are having. If one wants a relatively stable, reliable distributed computing project, seti-classic is out there, among others.

But someone has to help work the kinks out of new projects. Many times problems of scale don't show themselves unless and until one scales up the project.

On the other hand, I do sometimes feel we don't quite get the respect and acknowledgement appropriate to our contribution. .....<snip>

I don't appreciate it when project managers don't immediately let me know-- even via their web pages-- what is up, why things aren't working, when they might be working again, and what they need me to do to help them, so that I can figure out how best to use my time. The best non-profits respect their volunteer workers better than this. They wouldn't have volunteers if they didn't. And I suspect Seti@home is losing volunteers at a wholesale rate to other projects.

That is why some of us have become such big fans of Predictor@Home. Their staff do a really good job of keeping participants informed, and doing everything in their power to to keep the project going, and the credits coming.

[mikkyo]

Originally posted by ditchhook:
<B>Don't do it, iBook cools by the keyboard. If you do that for a long time, it'll burn.</B>
Thanks, I didn't realize that.

*snip*

While testing BOINC/climateprediction since yesterday, I noticed it crashed on me once in a 24 hour period, so I would like to create a "boinctab" routine that also restarts BOINC once an hour. Anyone done something like that in the past?

Closing your keyboard should put the machine to sleep. It should only keep running if you hacked it to do so.

As for auto restarting, yeah we have done stuff like that before.
The Ubero scripts were the most advanced, checking if the clients were running and if not restarting them. The Ubero script also setup memory sizes and number of clients based on the amount of RAM and number of CPUs you had. Too bad Ubero went into useless land.

You can make a script that starts boinc for you, and saves the process ID in a file in say /tmp. Then you make a script to check every hour and see if the process is still running by doing a "ps -p PID | grep Boinc", where the PID is the number from the file from the starter script and the Boinc is the partial name of your boinc client.
If you don't get anything back from that command, you restart the client using your starter script. If you do get something back from the grep command you exit.

Something like:
------Starter-----
#!/bin/sh
# remove old pid file if it is there
rm "/tmp/boincpid"
# cd boinc dir
cd "/Users/you/Boinc4"
# Start boinc in background
/usr/bin/nohup ./boinc_4.06_ppcG5 &
# Record PID ($! is the PID of the last process run
echo $! > "/tmp/boincpid"

----------------
-----Checker----
#!/bin/sh
# see if boincpid exist if not launch starter
echo "Checking if boinc is running..."
if [ ! -e /tmp/boincpid ]; then
echo "no boincpid, calling Starter"
cd /Users/you/Boinc4
./Starter
exit
fi
# Get PID from boincpid file and store in myPID
echo "Checking boincpid"
myPID=`cat "/tmp/boincpid"`
# Using PID see if process is running
isRunning=`ps -p $myPID | grep boinc`
# If not running launch
if [ ! $isRunning ]; then
echo "boinc not running, calling Starter"
cd /Users/you/Boinc4
./Starter
exit
fi
echo "boinc running, check complete"

----------------
Make sure there is a blank line at the end of each file.
Save each script as a plain text file without the .txt extension, then make them executable with
chmod a+rx Starter
chmod a+rx Checker
Put Both scripts in your Boinc4 folder(or whatever you call it, just make sure the paths match in the script.)
Then have your boinctab be something like:
-----Boinctab----
# At restart launch boinc in the background
@restart cd /Users/you/Boinc4; ./Starter
# Every 15 minutes of every hour of every day check for boinc running
0,15,30,45 * * * * cd /Users/you/Boinc4; ./Checker
----------------
Do a
crontab boinctab
and you are off and running.

The only problem with this kind of approach is that sometimes Boinc will crash and leave the workers running orphaned, potentially corrupting your work units. Also if the workers are still running and you launch boinc again, it could corrupt the workunits as well.
The only way around this is to do something fancy like check if any of the workers of any of the boinc clients are running, and kill them off. That would be a bit of a pain to write, so Ill leave that up to the daring.

Everything in /tmp is deleted at reboot, so the old boincpid will go away with a restart.

The only other potentially bad thing about this is that a user, you, could run the starter say 2 times in a row, and it would try and start boinc twice. I think boinc has lockfiles to prevent this, but I don't know for sure. You could get around this, by adding the Checker routines to the beginning of the Starter, before it removes the boincpid file. Then if a boincpid file existed when the start was run, it would make sure the boinc client wasn't running before removing it. If not it would just exit. Of course, you can just copy and paste most of the checker into the Starter, then you wouldn't need a separate checker, which is how our Ubero script was set up. It is just easier to explain things with the 2 separate scripts.

I tested the 2 scripts on my machine and they even work as coded, so enjoy.

[mikkyo]

Several macs have finally picked up Climate Prediction working models now.
THough Several other continue to have problems.
It seems like some models are fixed and others are still broken.

[Shaktai]

I got this e-mail from them. It confirms a problem with the Mac work units.

....in case you were trying to run climateprediction.net on a Mac and having trouble running the program (most likely within the past week), we have identified and corrected a bug with recent workunits.

So if you "Detach" and re-"Attach" to climateprediction.net using your CPDN/BOINC authenticator key, it should hopefully be OK now.