1) Message boards : Number crunching : ATLAS tasks dying like flies (Message 6533)
Posted 11 Aug 2019 by Thund3rb1rd
Post:
Three days ago, I started getting ATLAS tasks. They have all died a few seconds after starting. I'm not sure where to start looking. Suggestions would be most welcome.

I suspect a certain amount of this sort of thing is normal in a development account, but all of them???!!!

The machine is an Intel i7 using Windows 10, BOINC 7.14.2 and VM 6.0.10. Pretty plain stuff.

If this message should have gone into the ATLAS area, pls advise. I'm old and get confused easily.

Thanks.
2) Message boards : Number crunching : NEVER MIND: Can't attach to lcatdev on new Win 10 computer (Message 6515)
Posted 4 Aug 2019 by Thund3rb1rd
Post:
I'm having trouble attaching to my -dev account on my new win 10 machine. BOINC keeps telling me it isn't a BOINC project and stalks off with it's snooty little nose in the air. I know I'm doing something wrong. Anyone have a suggestion?

Edit: Never mind. Figured it out. God, I hate computers! They always do exactly what you tell them to do, even if it's stupid.
3) Message boards : Number crunching : An Alternative Approach For VM applications (Message 5560)
Posted 4 Oct 2018 by Thund3rb1rd
Post:
I don't/won't run LINUX so this great idea sounds like a really efficient way to lose volunteers like me. I've already trashed CPDN because it has become an unstable mess and no one over there seems to be able to fix it. Looks to me like LHC is reading their roadmap to oblivion.

Why is it that projects reach a point of reliable stability, then someone gets a wild hare to start dinking around with the architecture?

I am really sorry to read you will start messing with Windows after this LINUX thing is in place.
4) Message boards : Number crunching : One computer gets no work; the other gets it all (Message 5435)
Posted 16 Jun 2018 by Thund3rb1rd
Post:
First, thanks for your help. I appreciate it.

After more digging, I have a more precise definition of the situation. Both LHC and LHC-dev are under the mistaken impression that VM isn't installed on the one machine, which is the reason neither will give up work.

6/16/2018 8:45:12 AM | LHC@home | Sending scheduler request: Requested by project.
6/16/2018 8:45:12 AM | LHC@home | Requesting new tasks for CPU
6/16/2018 8:45:16 AM | LHC@home | Scheduler request completed: got 0 new tasks
6/16/2018 8:45:16 AM | LHC@home | No tasks sent
6/16/2018 8:45:16 AM | LHC@home | Message from server: VirtualBox is not installed
6/16/2018 8:45:16 AM | LHC@home | Tasks for AMD/ATI GPU are available, but your preferences are set to not accept them

6/16/2018 8:42:22 AM | lhcathome-dev | Sending scheduler request: Requested by user.
6/16/2018 8:42:22 AM | lhcathome-dev | Requesting new tasks for CPU
6/16/2018 8:42:25 AM | lhcathome-dev | Scheduler request completed: got 0 new tasks
6/16/2018 8:42:25 AM | lhcathome-dev | No tasks sent
6/16/2018 8:42:25 AM | lhcathome-dev | Message from server: VirtualBox is not installed
6/16/2018 8:42:25 AM | lhcathome-dev | Tasks for AMD/ATI GPU are available, but your preferences are set to not accept them

Of course it is. BOINC 7.10.2 and VM 5.2.12 including the 5.2.12 expansion pack have been installed on both boxes. Cosmology@home is running quite nicely on both machines using BOINC/VM.

So the question has become why do the two LHC-related projects "see" my VM install on one machine but not the other. The installations are identical, and work well on a non-LHC-related project?

As a final point, both LHC-related projects have run successfully on both machines in the past. I stopped allowing them to run on the one machine some time ago because of problems I was having with VM locking everything up. I decided to ease back into LHC/LHC-dev just recently... a decision I'm starting to regret.
5) Message boards : Number crunching : One computer gets no work; the other gets it all (Message 5433)
Posted 16 Jun 2018 by Thund3rb1rd
Post:
I've just noticed that one of my two computers hasn't gotten any work from LHCathome-dev for over a year. Granted, I set the 3 projects using VM to nowork, but that was just lately.

What would cause LHCathome-dev to send work to one of my machines, but not to the other?

Where should I start looking?
6) Message boards : Number crunching : What happened to the message boards? (Message 4164)
Posted 6 Oct 2016 by Thund3rb1rd
Post:
Hi, Laurence.

I saw that message, but at the risk of being obtuse, I don't understand what it has to do with the posts on the message boards for the past two months. There a handful of very recent posts, but only a handful and mostly from the past few days.

I switched over to the https:// URL a week or so ago, but only just now noticed the "missing" posts.

In the grand scheme of things, this isn't really that big a deal, I guess. At my age, I get confused easily.

So what am I missing?
7) Message boards : Number crunching : What happened to the message boards? (Message 4162)
Posted 6 Oct 2016 by Thund3rb1rd
Post:
I wanted to look something up from a post about app_config.xml and couldn't find it. Then I realized there are no posts dated later than the middle of August on either vLHCathome-dev or vLHCathome.

I didn't get that memo.
8) Message boards : News : Task and CPU limiter (Message 3921)
Posted 1 Aug 2016 by Thund3rb1rd
Post:
I've been successfully tweaking a combination of app_config and BOINC local settings to tailor my machines to what I want.

All of this really only affects three projects and the tasks don't take that long to run, so frankly, I'm not sure what problem is being addressed here.

As far as how much work is being downloaded, is that really an issue? Personally, I take what I can get and don't throw a fit.

Or am I so far off-base here that this posting is simply ludicrous?
9) Message boards : Number crunching : Virtual Machine Inaccessible (Message 3632)
Posted 4 Jul 2016 by Thund3rb1rd
Post:
Never mind, Ivan. I got it figured out. The task belonged to Cosmology@home. The stack dump indicated that the VM machine was deregistered, but it wasn't. I'll toss it over their fence and see what happened.

Thanks for your help.
10) Message boards : Number crunching : Virtual Machine Inaccessible (Message 3631)
Posted 4 Jul 2016 by Thund3rb1rd
Post:
Suggest you abort the task and see if the new one starts up OK.


Actually, that was my first of many guesses, but I don't have any such task. I have 3 vLHC tasks assigned - one from -dev and two from vLHC. All three tasks are actually running (according to BOINC). The VB Manager shows all three VM as being powered off but I've asked about that already. Apparently it doesn't matter.

Bottom line - I no longer have any such task; just the abandoned VM and it's apparently broken.

I'm going to raise an issue here in general terms since it may simply be my imagination, and if not, may not even apply to vLHC. After I got VB correctly installed and functional and was processing LHC stuff, I also started getting VB stuff from cosmology@home, which is okay - I wanted to. However, I'm starting to get the feeling that the two - LHC and Cosmology - are not playing well together. Nothing concrete - yet. Just a feeling.
11) Message boards : Number crunching : Virtual Machine Inaccessible (Message 3629)
Posted 4 Jul 2016 by Thund3rb1rd
Post:
One of my tasks has gotten screwed up somehow. Here is the data from the VB Manager screen:

Runtime error opening 'C:\ProgramData\BOINC\slots\1\boinc_a4ae6e89571c0a8a\boinc_a4ae6e89571c0a8a.vbox' for reading: -103(Path not found.).

F:\tinderbox\win-5.0\src\VBox\Main\src-server\MachineImpl.cpp[740] (long __cdecl Machine::i_registeredInit(void)).
Result Code: E_FAIL (0x80004005)
Component: MachineWrap
Interface: IMachine {f30138d4-e5ea-4b3a-8858-a059de4c93fd}

Not sure exactly what all this means, but the (Path not found) part doesn't sound good.
12) Message boards : Number crunching : Credit Issues (Message 3622)
Posted 1 Jul 2016 by Thund3rb1rd
Post:
This is a testing project.
Credits are not that important here.
The main concern is to get the apps working properly.

You missed the answer.

No I didn't. You are confusing the main concern of the project owner with the concern of a cruncher. As a cruncher, I can influence neither the project setup, nor its scientific value. I can only donate time and hardware. For credit, which costs the project owner nothing. And not only does it cost nothing, it does not in any way prevent or detract the project owner from getting the apps working properly. So how did you come up with the idea, that by not receiving credit you will somehow help the project?


Completely, totally in agreement with the sentiment expressed here by KPX. The ONLY satisfaction a cruncher has directly available are the credits. Without them, our effort is meaningless. From our standpoint, a project is a project is a project. Whether its a test environment or a production one is absolutely immaterial to our side of the fence. Our computers neither know the difference, nor do they (or we) care.

A bit of credit philosophy here. About five years ago, I began to notice a really significant difference in the credit awarded by the projects I was participating in. REALLY significant. So, I set up my own little experiment. For the past five years, I have done NOTHING to alter my production setup. At the end of five years, what was apparently a statistical anomaly became as bright as day. I participate in eight projects. Of these eight, two - Einstein@home and milky@home - are very liberal with credits. Two others - seti@home and climateprediction.net - are extremely stingy with credit. All eight projects receive exactly the same amount of computer time, and it's apparent that over the years, available work from each project has tended to even itself out, resulting in what I see as a viable picture of BOINC efforts.

In MHO, of course.
13) Message boards : Number crunching : VBox shows "Powered off" (Message 3621)
Posted 30 Jun 2016 by Thund3rb1rd
Post:
What does this refer to? BOINC shows the tasks as running, but the VBox manager shows them as being "Powered off". Worth being concerned about?
14) Message boards : Number crunching : Very long estimated run times for Theory Simulation v1.07 (Message 3599)
Posted 22 Jun 2016 by Thund3rb1rd
Post:
Thanks for the information.

You were right on. The task actually ran to completion in about 12 and a half hours and was validated.
15) Message boards : Number crunching : Very long estimated run times for Theory Simulation v1.07 (Message 3594)
Posted 22 Jun 2016 by Thund3rb1rd
Post:
As I write this, I have four vLHC-dev tasks in one of my machines - two CMS, one LHCb and one Theory Simulation v1.07.

The estimated run times for CMS and LHCb are about 13 hours and about 40 minutes, respectively.

However, the estimated run time for the TS task is just under 4 DAYS! Is this correct, or am I missing something?

I don't particularly care, mind you, but with VB's propensity to ignore BOINC run-time allocation management, the rest of my projects are going to be awfully unhappy.
16) Message boards : Number crunching : VBox issues (Message 3592)
Posted 21 Jun 2016 by Thund3rb1rd
Post:
Another thing. BOINC control over these tasks seems problematic at best. For example, I've figured out that I can replace the feeble BOINC run-time allocation by setting projects to and from "won't get new tasks" but even this is occasionally ignored.
17) Message boards : Number crunching : VBox issues (Message 3590)
Posted 21 Jun 2016 by Thund3rb1rd
Post:
Thanks, but I found it. Stumbled over it, actually, but found it...
18) Message boards : Number crunching : VBox issues (Message 3589)
Posted 20 Jun 2016 by Thund3rb1rd
Post:
Crystal Pellet - Where did you get the expansion pack for 5.0.12?

All I could find on the Oracle site was the one for 5.0.20.
19) Message boards : Number crunching : VBox issues (Message 3588)
Posted 20 Jun 2016 by Thund3rb1rd
Post:
Two things so far:

1) Setup seems incredibly complex. Too much dinking around with configuration files and balancing different versions of BOINC and VB with the operating system. I'm a plug-and-play sort of guy, so I find this all unnecessarily tedious and annoying. By the way, I get that this is a development environment, but still....

2) Once VB starts a task, it/they freeze out all other projects. BOINC doesn't seem to exercise any control over the run time allocation. I have three LHC2/LHC-dev tasks running right now that once they started, nothing else has gotten any time at all even on a quad-core machine.



©2020 CERN