Lawrence,

I mention HD usage because Sekerob's response about the problem was to check the minimum requirements of the projects, specifically the HD limit I had set in my prefs (200 MB). So I upped the limit to 2 GB. (as you see, it didn't help)

I'm using the latest recommended BOINC version for OS X (5.10.7). As I said in my first post, I'm using local prefs (defined in the Simple View of BOINCManager), and the limits are right now 2 GB HD, 30% CPU, don't work when on battery, work after 1 minute idle.

Anyway, HD free space has never decreased noticeably. When the problem reappeared last night, I had more than 2.7 GB free, and they are still free now.
So I can't find a link with the problem.



I think having a preferences-set limit of 2 GB, and 2.7 GB free on the HD, discards that possiblity.



Just for completeness, here I recap once again my prefs, local and otherwise:
Local: 2 GB, 30% CPU, 1 minute idle, no working while on battery
Web (are they used? who knows? not the normal user, that's for sure): no more than 50% CPU, no more than 75% of total mem, 3 mins idle, at most 1 CPU, switch apps every 120 mins, no more than .3 GB HD, at least .5 GB free HD, no more than 75% HD, write to disk at most every 99, stay in memory, connect to network every 3 days.




The log never reported more than the oh-so-useful "zero status, no file, may need to reset".




5.10.7 (latest recommended for OS X)



Interesting. The problem is, I already was using BOINC when I found WCG months ago, so I didn't even realize that they had a different BOINC recommendation.

The other (and I feel much bigger) part of the problem, of course, is the lack of any sanity checking - or at least, the lack of any reporting of the results of whatever sanity checking. "Hey, user, we'll try to run but you should be aware that this BOINC release is not an officially supported one by WCG; if something goes wrong, you should yadda yadda yadda".
EVEN IF I had gotten into BOINC thru WCG, I would have surely updated my BOINC installation as soon as I had discovered that a new version was available. Which is what normal users do, and in fact is what they are usually told to do. So sanity checking AND REPORTING gets mandatory - and its lack, crippling. Which is the case.



Shame on them. And shame on WCG for not letting me know without me having to go out of my way.



As said, 2.7 GB free, 2 GB prefs-limited. So: check.



Check.



Check.



There is no VM setting in OS X that I can think of, and anyway, I set 75% of memory in the web's prefs page. So, check.



I have tried that, and anyway I guess that the effects of that are a subset of what a project reset does, which I did four times yesterday. So... check, and Not Working For Me.



I do feel that this problem asks for some upsetting :P. Someone, somewhere, is not doing his/her homework.

Just for the record: I'm keeping WCG suspended on my computer, as nothing new has been suggested that could improve the situation, and WCG-induced CPU maxing-out has managed to blaze a direct trail for getting into my nerves :P.
(and I'm not downgrading BOINC, since anyway good old SETI@Home is happily crunching away).

From the early days of 5.10 CA's (Community Advisors) have been vocal on WCG not recommending 5.10. The WCG download page now explicitly states the version it offers:

We can't win them all, even if going out of our way to help users going. Someone with advanced MAC experience will be able to tell how to address the VM side. Here's a oogled post of someone even putting his swapfile on a dedicated partition and commenting on a 80mb default: http://www.xlr8yourmac.com/OSX/OS_x_swapfile_sizes.html
http://text.dslreports.com/faq/allthingsmac/6_OSX

strange thing is, i think you are the first to report issues that are linked to disk space on a Mac (I'm not sure now if it is the problem, but we can only work from eliminating possibilities). That OS with it's strong GUI is probably less needing looking under the hood.

I see that WCG, in its downloads page, offers a 5.8.x version.
I can't see anywhere else (except in these forums, that is) any remote hint that the user SHOULD NOT update to the latest officially recommended BOINC version, as stated in the official BOINC homepage.
How exactly am I supposed to know that?
This should be made EXPLICIT, in-your-face, multiple times and in many ways: in the WCG info web pages, the WCG prefs web pages, and the BOINC logs.



Oh, please.
Just for the record, in case some Mac OS X user comes here and sees these posts:
THERE IS NO REASON to touch your VM, except in VERY specific cases. In the earliest versions of OS X, it could SOMETIMES make sense. But it's been a long time (OS X 10.2 era) since I last saw anyone talking about swap partitions and the like for OS X.
By the way, the first page you link to confirm that: they refer to OS X 10.1 (circa 2001).
(the second link doesn't even mention virtual memory)

In case you want some more context, my OS X with it standard VM settings (standard because there is no official way to change them) is using right now a 3GB swap file, of which more than 2.5 GB are reported as in use by the OS. OS X grows and shrinks its swap files as needed; it starts with few MB at boot and I've seen it grow past 5 GB and back under 2GB as the OS sees fit.
For completeness sake again, I have 3 GB RAM, in a Macbook Core2Duo. I have right now 2.7 GB of free HD space. When freshly booted, the HD free space is about 6 GB; the VM grows in that free space.



That's it, we indeed can't be sure if that is the problem.

And since there is absolutely no hint about what could exactly be the problem, any user with remotely similar symptoms could be having it.

And as I said, I have been suffering this problem for weeks before coming here to look for some solution. Other users could just be suffering the CPU peaks and don't
realize, or don't bother to look for the reason, or don't know how to! Because, as said, there is no hint, even in BOINC's logs, that HPF2 is causing CPU extreme usage. Only that "it exited with no files". Useful!

The BOINC CPU throttle of 30% means it runs 3 seconds at 100% and pauses for 7 seconds.... on all projects, both 5.8 and 5.10. At some point the throttling itself caused problems in the science, with the only way avoiding offered was to set it at 100%.

Like Didactylos already mentioned, during checkpointing the throttle is ignored as disk i/o takes priority.

As for the original message:

Task lf432_00028_11 exited with zero status but no 'finished' file
If this happens repeatedly you may need to reset the project.

One known cause is the system clock being synchronized by the system. If that happens middle of the science, particularly backward, BOINC considers a timekeeping problem and tries to return a project to the last checkpoint. Just wonder if the throttling alternation 3:7 and the system time keeping might have something to do with this. Very few i've heard of set it to 30%, considering that the processes use pure idsle time and are at lowest priority i.e. even at 100% should hardly impair use. Full tilt is known to cause the fan to run all the time, but another Mac user reported that 90% was enough for that not to happen.