World Community Grid - View Thread - errors after converting my VMs to BOINC

World Community Grid Forums

Category: Support

Forum: BOINC Agent Support

Thread: errors after converting my VMs to BOINC

Quick Go »

No member browsing this thread

Thread Status: Active
Total posts in this thread: 6

[ ]

Author

This topic has been viewed 613 times and has 5 replies

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


errors after converting my VMs to BOINC

I recently converted about 20 virtual machines from the UD agent to BOINC. All results are returning as invalid. These VMs have been running fine under UD for over two years. I only noticed the issue after I dropped 20 ranks in two days. :) The VM agents are installed at c:\program files\BOINC.

Here is one of the erorr logs:

<core_client_version>5.10.30</core_client_version>
<![CDATA[
<message>
The system cannot find the path specified. (0x3) - exit code 3 (0x3)
</message>
<stderr_txt>
World Community Grid AutoDock (projects/www.worldcommunitygrid.org/wcg_faah_autodock_5.42_windows_intelx86) version Failed to get VersionInfo size: 1812

Failed to get VersionInfo size: 1812
INFO: No state to restore. Start from the beginning.
ERROR: Restoring checkpoint failed. Unable to restore state!
INFO:[21:32:57] Start AutoGrid...

autogrid: autogrid4: Successful Completion.
INFO:[21:41:11] End AutoGrid...
Beginning AutoDock...
INFO: Setting num_generations: 27000
Setting maxGen to 6750
call_glss(): pop_size: 200 num_evals: 1000000 start: [21:43:40]
call_glss(): begin 2nd for loop over pop: [21:43:40]
call_glss(): end 2nd for loop over pop: [21:43:40]
call_glss(): end 1st do/while: [21:43:40]
call_glss(): end: [21:47:38]
_maxGenSeenSoFar changed: 6750
call_glss(): pop_size: 200 num_evals: 1000000 start: [21:47:38]
call_glss(): begin 2nd for loop over pop: [21:47:39]
call_glss(): end 2nd for loop over pop: [21:47:39]
call_glss(): end 1st do/while: [21:47:39]
call_glss(): end: [21:51:36]
About to enter main loop...(dockings already completed: 0)
call_glss(): pop_size: 200 num_evals: 10000000 start: [21:51:36]
call_glss(): begin 2nd for loop over pop: [21:51:36]
call_glss(): end 2nd for loop over pop: [21:51:36]
call_glss(): end 1st do/while: [21:51:36]
call_glss(): end: [22:38:41]
Updating Best Energy for WU: 0.00
call_glss(): pop_size: 200 num_evals: 10000000 start: [22:39:34]
call_glss(): begin 2nd for loop over pop: [22:39:34]
call_glss(): end 2nd for loop over pop: [22:39:34]
call_glss(): end 1st do/while: [22:39:34]
call_glss(): end: [23:27:25]
Updating Best Energy for WU: -4.35
call_glss(): pop_size: 200 num_evals: 10000000 start: [23:27:48]
call_glss(): begin 2nd for loop over pop: [23:27:48]
call_glss(): end 2nd for loop over pop: [23:27:48]
call_glss(): end 1st do/while: [23:27:48]
call_glss(): end: [00:17:28]
call_glss(): pop_size: 200 num_evals: 10000000 start: [00:17:44]
call_glss(): begin 2nd for loop over pop: [00:17:44]
call_glss(): end 2nd for loop over pop: [00:17:44]
call_glss(): end 1st do/while: [00:17:44]
call_glss(): end: [01:00:09]
call_glss(): pop_size: 200 num_evals: 10000000 start: [01:00:27]
call_glss(): begin 2nd for loop over pop: [01:00:27]
call_glss(): end 2nd for loop over pop: [01:00:27]
call_glss(): end 1st do/while: [01:00:27]
call_glss(): end: [01:49:08]
call_glss(): pop_size: 200 num_evals: 10000000 start: [01:50:03]
call_glss(): begin 2nd for loop over pop: [01:50:03]
call_glss(): end 2nd for loop over pop: [01:50:03]
call_glss(): end 1st do/while: [01:50:03]
call_glss(): end: [02:39:17]
Updating Best Energy for WU: -4.59
World Community Grid AutoDock (projects/www.worldcommunitygrid.org/wcg_faah_autodock_5.42_windows_intelx86) version Failed to get VersionInfo size: 1812

Failed to get VersionInfo size: 1812
INFO: No state to restore. Start from the beginning.
ERROR: Restoring checkpoint failed. Unable to restore state!
INFO:[02:42:43] Start AutoGrid...

autogrid: autogrid4: Successful Completion.
INFO:[02:54:09] End AutoGrid...
Beginning AutoDock...
INFO: Setting num_generations: 27000
Setting maxGen to 6750
call_glss(): pop_size: 200 num_evals: 1000000 start: [02:56:19]
call_glss(): begin 2nd for loop over pop: [02:56:19]
call_glss(): end 2nd for loop over pop: [02:56:19]
call_glss(): end 1st do/while: [02:56:19]
call_glss(): end: [03:00:44]
_maxGenSeenSoFar changed: 6750
call_glss(): pop_size: 200 num_evals: 1000000 start: [03:00:44]
call_glss(): begin 2nd for loop over pop: [03:00:44]
call_glss(): end 2nd for loop over pop: [03:00:44]
call_glss(): end 1st do/while: [03:00:44]
call_glss(): end: [03:05:48]
About to enter main loop...(dockings already completed: 0)
call_glss(): pop_size: 200 num_evals: 10000000 start: [03:05:48]
call_glss(): begin 2nd for loop over pop: [03:05:48]
call_glss(): end 2nd for loop over pop: [03:05:48]
call_glss(): end 1st do/while: [03:05:48]
call_glss(): end: [03:54:54]
Updating Best Energy for WU: 0.00
[ERROR] Failed to write to wcg_checkpoint_11.ckp during copy operation. Error: 2
World Community Grid AutoDock (projects/www.worldcommunitygrid.org/wcg_faah_autodock_5.42_windows_intelx86) version Failed to get VersionInfo size: 1812

Failed to get VersionInfo size: 1812
INFO: No state to restore. Start from the beginning.
ERROR: Restoring checkpoint failed. Unable to restore state!
INFO:[03:58:14] Start AutoGrid...

autogrid: autogrid4: Successful Completion.
INFO:[04:12:40] End AutoGrid...

</stderr_txt>

[Mar 7, 2008 8:58:45 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: errors after converting my VMs to BOINC

It sounds like a permissions issue. What user are you running BOINC as? Does it have full control of the BOINC directory?

Can you explain more about your VM setup?

[Mar 7, 2008 9:11:13 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: errors after converting my VMs to BOINC

Ahh, I installed as a service with .\administrator as the service account but then changed them all to run from the local system account. Running under teh system account is still my preference. I was previously running the UD agent as a service using srvany. These are XP SP2 VMs running under vmware Workstation 5.5 on a eight core server.

[Mar 7, 2008 9:36:24 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: errors after converting my VMs to BOINC

You sound like you know what you're doing - just change the directory and file permissions appropriately.

However, one detail: BOINC runs fine on multiple cores. You don't need to use VMs to get maximum performance out of your computer. This is one of the major advantages of using BOINC.

[Mar 7, 2008 9:45:25 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: errors after converting my VMs to BOINC

Thanks. I hadn't considered this -- maximizing the workload was one of the drivers for running the UD agent under VM. That said, I do have the BOINC agent running on the host server as well, and crunching performance seems marginal even without the VMs running. What can I tweak to maximize crunching performance without killing interactive process performance (I also use this "server" is used as a general-purpose workstation)?

Thanks for the help. I was pretty confortable with the UD agent but haven't used BOINC since the early SETI@home days.

[Mar 7, 2008 9:59:30 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: errors after converting my VMs to BOINC

Make sure your preferences allow BOINC to use all 8 cores. Also, make sure the CPU percentage is set to 100%, and the memory settings are generous.

Once you have done this, update BOINC and wait for it to download enough work to fill all the cores (don't ask, it's a bit unintuitive).

As for other possible performance issues - the only thing I can think of is that your VMs may be interfering even if they aren't actually running anything. Because the VM runs at normal priority, the idle process on the VM has higher priority than BOINC on the host machine. I think. That may depend on the particular VM you use.

[Mar 7, 2008 10:11:22 PM]

[ ]