Errors on wu after 48 minutes:
196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED

13 valid so far, no invalids. They run about 9 minutes each on 7 cores of an i7-4770 (Win7 64-bit, BOINC 7.6.9, VBox 5.0.3r102449).

Thanks everyone, please keep comments coming!

I will be out of touch until about Sept 9th, as I'm moving overseas.

I'll make sure to catch up on messages left here when I'm back. A number of things are fixed, but not live yet, including suspend/resume improvements, file deletion issues, app preferences... These will go live once I get back to work around the 9th, I'd rather do that then do it now and have there be issues that I can't deal with. Talk to you all later and thanks for the help!

I was keen to try this out for you so I installed the latest Boinc version 7.6.9, which comes with VirtualBox version 4.3.12
I appear to have downloaded all the needed files, but my 2 work units won't start; there is the message in the Event Log:

04/09/2015 11:53:02 | cosmohome | task postponed 86400.000000 sec: Please upgrade BOINC to the latest version.

EDIT: Ignore this message. I stopped, then restarted my Boinc client, and now the work units appear to be running correctly.

MORE EDIT: My work units are running in the Boinc Manager, and also in the VirtualBox Manager, but there is no CPU activity, and the work units are still sitting at 0.1% after 30 minutes running.

---

The biggest threat to public safety and security is not terrorism, it is Government abuse of authority.
Bitcoin Donations: 1Le52kWoLz42fjfappoBmyg73oyvejKBR3

The best I've been able to find out why my work units aren't running properly is this entry in the Virtualbox log:

00:00:00.355858 WARNING! 64-bit guest type selected but the host CPU does NOT support HW virtualization.

My CPU is an Intel Pentium 4 640 @ 3.2GHz. It is 64-bit.

I guess I'll have to give up on testing out this project. Pity.

---

The biggest threat to public safety and security is not terrorism, it is Government abuse of authority.
Bitcoin Donations: 1Le52kWoLz42fjfappoBmyg73oyvejKBR3

My CPU is an Intel Pentium 4 640 @ 3.2GHz. It is 64-bit.

I think your CPU supports virtualization, you probably just need to enable it in the motherboard BIOS.

I think your CPU supports virtualization, you probably just need to enable it in the motherboard BIOS.

If you are referring to VT-x then according to CPU-Z my old Pentium 4 processor doesn't support it.

---

The biggest threat to public safety and security is not terrorism, it is Government abuse of authority.
Bitcoin Donations: 1Le52kWoLz42fjfappoBmyg73oyvejKBR3

If you are referring to VT-x then according to CPU-Z my old Pentium 4 processor doesn't support it.

That might be because it is not enabled in the motherboard. (It is an easy Google search for the CPU.)

My search on the Intel website:

http://ark.intel.com/products/27480/Intel-Pentium-4-Processor-640-supporting-HT-Technology-2M-Cache-3_20-GHz-800-MHz-FSB

revealed: Intel® Virtualization Technology (VT-x) ‡ No

So I guess I can't use this computer for Beta testing.

---

The biggest threat to public safety and security is not terrorism, it is Government abuse of authority.
Bitcoin Donations: 1Le52kWoLz42fjfappoBmyg73oyvejKBR3

They should know. If the option is not available in the BIOS, then it won't be available in any case.

Having got this project working properly on my (other) 64-bit Windows computer, I am keen to test it out on my 64-bit Linux machine.
It runs Ubuntu 14.04.3 and has Boinc version 7.2.42 installed.
I have downloaded virtualbox-5.0_5.0.2-102096~Ubuntu~trusty_amd64.deb, but I haven't installed it yet.
Will this combination work OK on this project, and is there anything special that I need to do to configure the VirtualBox installation, or will the work units just work fine without doing anything else?

---

The biggest threat to public safety and security is not terrorism, it is Government abuse of authority.
Bitcoin Donations: 1Le52kWoLz42fjfappoBmyg73oyvejKBR3

Update: I've got some work on my Linux machine, and I've finished 2 work units, but it's hard going!
The work units stop after about 20 minutes with no reason given in the Event Log, but in the work unit Properties it notes:

State Waiting to run (Scheduler wait: VM job unmanageable, restarting later.)

When I reboot the computer the work unit picks up where it left off, except that the Elapsed time drops down to the last checkpoint time.

Both work units stopped again before they finished, needing another reboot.

I'll finish my current batch of work, but I'm not doing any more Linux work until I've figured out what I did wrong.

---

The biggest threat to public safety and security is not terrorism, it is Government abuse of authority.
Bitcoin Donations: 1Le52kWoLz42fjfappoBmyg73oyvejKBR3

Hello,

Sorry if question already asked somewhere else.

Running camb_boinc2docker on a 4770K HT ON, but although BM indicates "8 CPU", it uses only 16 % of it ?

How comes it uses all cores with a so small "intensity" ?

Thank You

Phil1966

Running camb_boinc2docker on a 4770K HT ON, but although BM indicates "8 CPU", it uses only 16 % of it ?

How, and more importantly when did you check it? The CPU usage starts out quite low at the beginning of a run, but increases towards the end, with a lot of variation in between. On my i7-4770, running on 7 cores, BOINCTasks shows the "CPU %" at 99+% at the end of the run. And Task Manager shows over 90% most of the time. But they are much less, only a few percent, at the start.

I am using BOINC 7.6.9 and VBox 5.0.4 on a Win7 64-bit machine.

Hey guys, OK, I'm mostly back, starting to get back into things. Thanks everyone for your patience.


M0CZY: I think you guys have it right. Unfortunately you need VT-x or AMD-v (and it needs to be enabled in the BIOS, which is also unfortunately sometimes not the default setting) for Virtualbox to run 64bit guests. Sounds like your one computer would only work for camb_legacy. As for your Linux computer, the problem seems to be related to checkpointing. Every time a checkpoint is made, the old one should be deleted, but for some reason it looks like you're getting an error there, even on jobs which you eventually complete successfully, e.g. here. For the moment I'm going to update the way checkpointing is done anyway, so let me hold off on trying to fix this right now and see if that just fixes it. Thanks for you comments!


Phil1966: I echo Jim1348's response, are you waiting a little while to check? At the very beginning of the job we are booting up into a lightweight operating system within the VM, and much of this code is single threaded. After at most maybe 1min you should see everything go full throttle. If you never see it go full throttle, that could be a problem, let me know.

When the percentage of the job reaches about 46,4 execution stops with an error:

task postponed 86400.000000 sec: VM Hypervisor failed to enter an online state in a timely fashion.

What can I do with this?

When the percentage of the job reaches about 46,4 execution stops with an error:

task postponed 86400.000000 sec: VM Hypervisor failed to enter an online state in a timely fashion.

What can I do with this?

I assume it was this result? Unfortunately the log is chopped off at 64 kilobytes and so I can't see what happened right before it looks like your vboxheadless.exe went down. Has this happened again, I only see that one result for the moment? Does rebooting your computer or upgrading the version of Virtualbox (to the latest 4.X.X or even 5) resolve it? I'm also curious if the VM ever made it to the actual computation in the first place. If it did, you should have seen your CPU usage jump to 100% (or whatever max you have set in your BOINC preferences). Did you notice if that happened?

The error occured due to disabled in the BIOS Virtualization Technology

After Virtualization Technology in BIOS is "ON", application work properly

Thanks!

Thanks for tracking it down and reporting back. A couple of people seem to have had issues with this VT-x thing. It would be better if we autodetected hosts without VT-x and never even tried to run these jobs on them, and I think in theory that's possible. It may even be that it already is trying to do so, but failing, at least in your cases. I will look into it.

Just to inform you that when calculating the job only on a one core CPU (with the use of app_config.xml), it fails with the error about exceeding the permissible time calculation.

For example:

http://beta.cosmologyathome.org/result.php?resultid=10517

Run time 39 min. 32 sec.
CPU time 38 min. 25 sec.

<message>
exceeded elapsed time limit 2357.33 (86400.00G/36.65G)
</message>

When calculating jobs on two cores and more error is missing.

For example:

http://beta.cosmologyathome.org/result.php?resultid=10600

Run time 25 min. 46 sec.
CPU time 49 min. 8 sec.

My computer (259482):

http://beta.cosmologyathome.org/show_host_detail.php?hostid=259482