1) Forums : Technical Support : script_validator (camb) is down (Message 22009)
Posted 28 Nov 2018 by Jesse Viviano
Post:
Is the camb_legacy validator intentionally shut down in preparation for the server upgrade, or is it supposed to be running? It appears to have been down for over a week because I still have camb_legacy tasks that were finished on or after November 19, 2018 that are still waiting for validation.
2) Forums : Technical Support : script_validator (camb) down (Message 21965)
Posted 14 Oct 2018 by Jesse Viviano
Post:
It is down as of this writing. I am posting this here because the same validator appears to have been started up again, but has quit working again.
3) Forums : Technical Support : VM Hypervisor (Message 21713)
Posted 4 Jun 2018 by Jesse Viviano
Post:
When I see this status, the best way to clear it is to quit the BOINC manager using the Exit BOINC command in the File menu, ensure that the check box for "Stop running tasks when exiting the BOINC manager" is checked, click "OK", and then launch the BOINC manager again. This will force BOINC to try to restart the stuck work unit's virtual machine, and it usually does get the virtual machine working again.
4) Forums : Technical Support : Server Out of Disk Space (Message 21688)
Posted 25 May 2018 by Jesse Viviano
Post:
Others are still having trouble uploading results. Right now is not the time for additional work to be generated because that will make the problem worse. The upload server obviously needs to have fewer files on its drive to open up space for more files to be uploaded. For that to happen, people must report any completed and completely uploaded work so that their corresponding files can be validated, assimilated, and then deleted for the server to make room for more files to be uploaded.

Crunchers can help by updating the project if they have any results marked as "Computation error" or "Ready to report" so that their files can be processed, assimilated, and deleted to make room for more files awaiting upload.
5) Forums : Technical Support : DNS failure regarding cosmologyathome.org and www.cosmologyathome.org (Message 21669)
Posted 20 Mar 2018 by Jesse Viviano
Post:
There are two ways to fix this. First is to reboot, forcing the DNS resolver cache to be purged. The second way is to launch either PowerShell or the classic Windows command prompt also known as cmd.exe in administrator mode and then enter the command "ipconfig /flushdns" to manually flush the DNS cache.
6) Forums : Technical Support : DNS failure regarding cosmologyathome.org and www.cosmologyathome.org (Message 21666)
Posted 18 Mar 2018 by Jesse Viviano
Post:
Your domains cosmologyathome.org and www.cosmologyathome.org no longer point to your server's IP address of 194.57.221.140, but to some page that Gandi.net says is you parking your domain. Please get your DNS fixed. To reach this site and to upload and report my tasks, I had to create static entries in my home broadband router's DNS server pointing those two domains to 194.57.221.140.
7) Forums : Technical Support : postponed VM hypervisor failed to enter and online state in a timely fashion (Message 21665)
Posted 18 Mar 2018 by Jesse Viviano
Post:
This error goes away for me when I shut down BOINC and all of the computing apps, and then restart it.
8) Forums : Technical Support : camb_boinc2docker fails after 15 minutes (Message 21498)
Posted 18 Aug 2017 by Jesse Viviano
Post:
Go to the project's directory. On Windows 10, it will be located at C:\ProgramData\BOINC\projects\www.cosmologyathome.org. Then create a file named "app_config.xml" and put the following code into it:
<app_config>
    <app_version>
        <app_name>camb_boinc2docker</app_name>
        <plan_class>vbox64_mt</plan_class>
        <avg_ncpus>8</avg_ncpus>
    </app_version>
</app_config>


This will force the Cosmology@home application to use no more than 8 virtual CPUs, or vCPUs for short. There is a bug in Cosmology@home where it will crash if it uses more than 8 vCPUs per task. I noticed that your Ryzen machine has 6 physical cores, so it can provide up to 12 vCPUs thanks to its SMT capability. The project's administrators should configure the project to use no more than 8 vCPUs to work around this bug.
9) Forums : Technical Support : Virtualized tasks monopolize all cores in BOINC but only use around 40% of the CPU (Message 21497)
Posted 18 Aug 2017 by Jesse Viviano
Post:
I was experimenting around again, and found that 9 vCPUs would cause tasks to crash. I guess that the project should configure its VMs to use a maximum of 8 vCPUs to keep tasks from crashing instead of relying on users to use an app_config.xml file to set this up.
10) Forums : Technical Support : Virtualized tasks monopolize all cores in BOINC but only use around 40% of the CPU (Message 21266)
Posted 31 Dec 2016 by Jesse Viviano
Post:
Experiments where I used the app_config.xml to assign 12 vCPUs to each task failed and had to be aborted.
11) Forums : Technical Support : Virtualized tasks monopolize all cores in BOINC but only use around 40% of the CPU (Message 21260)
Posted 22 Dec 2016 by Jesse Viviano
Post:
I have done some experimenting, and found the following: using all 24 vCPUs left more than half of the virtual cores idle when the task does not hang. (The first task I try to execute with 24 vCPUs always hangs during the booting of the virtual machine and needs to be aborted. The rest go through fine.) Switching to 9 vCPUs per task allowed my CPU to be fully utilized when the other logical cores are filled with single core tasks, but there was hardly any speedup from using 8 vCPUs per task. Perhaps the other single core tasks competed for my CPU's memory controllers enough to slow down the multicore tasks. Giving tasks 10 vCPUs would cause them to consistently hang during the virtual machine's boot sequence, forcing me to abort them. I haven't had the time to try 11 or 12 vCPUs per task yet.
12) Forums : Technical Support : HTTPS site is down (Message 21208)
Posted 24 Nov 2016 by Jesse Viviano
Post:
The HTTPS site is now running again, so this can be crossed off the list of things to do to recover the project from last weekend's crash.
13) Forums : Technical Support : HTTPS site is down (Message 21207)
Posted 21 Nov 2016 by Jesse Viviano
Post:
I know that you have more pressing matters at this time (the project is recovering from crashing over the weekend and as of the time of this writing the feeder and the transitioner are down), but I wanted to let you know that the HTTPS version of this website is down. Please fix this after you have finished recovering from the crash and fixed the feeder and transitioner.
14) Forums : Technical Support : Virtualized tasks monopolize all cores in BOINC but only use around 40% of the CPU (Message 21203)
Posted 5 Nov 2016 by Jesse Viviano
Post:
I have cleaned out some unnecessary lines in my app_config.xml. A better one is below:
<app_config>
    <app_version>
        <app_name>camb_boinc2docker</app_name>
        <plan_class>vbox64_mt</plan_class>
        <avg_ncpus>8</avg_ncpus>
    </app_version>
</app_config>
15) Forums : Technical Support : Computing while skipped (Message 21200)
Posted 30 Oct 2016 by Jesse Viviano
Post:
You might want to upgrade your version of BOINC. BOINC version 7.6.22 is released to general production.
16) Forums : Technical Support : Lingering task (Message 21197)
Posted 29 Oct 2016 by Jesse Viviano
Post:
Abort the task. Someone else has completed the replacement task, so you will never get credit for the work your computer did. I have never run into a never-ending task problem with Cosmology@home, so it might be a one-off.
17) Forums : Technical Support : Virtualized tasks monopolize all cores in BOINC but only use around 40% of the CPU (Message 21195)
Posted 26 Oct 2016 by Jesse Viviano
Post:
I found that limiting the number of vCPUs to 8 per task and not limiting the number of tasks run at once maximizes my speed and CPU utilization. (24 logical cores / 8 vCPUs per task gives you 3 tasks running at once.) HyperThreading is not the problem I am seeing. I think that something is wrong within either VirtualBox or the Linux scheduler in each virtual machine to prevent it from maxing out more than 8 vCPUs per virtual machine based on my evidence, the ATLAS@home evidence, and previous posts by other users in this thread. Any time I have given a Linux VM more than 8 vCPUs, it does not max them out.

Here is the app_config.xml that I have used that maxes out my CPU:
<app_config>
    <app>
        <name>camb_boinc2docker</name>
    </app>
    <app_version>
        <app_name>camb_boinc2docker</app_name>
        <plan_class>vbox64_mt</plan_class>
        <avg_ncpus>8</avg_ncpus>
    </app_version>
</app_config>
18) Forums : Technical Support : Virtualized tasks monopolize all cores in BOINC but only use around 40% of the CPU (Message 21193)
Posted 25 Oct 2016 by Jesse Viviano
Post:
ATLAS@Home has the same problem that Cosmology@home does when dealing with more than 8 vCPUs per virtual machine. Please see http://atlasathome.cern.ch/forum_thread.php?id=562, http://atlasathome.cern.ch/forum_thread.php?id=568, and http://atlasathome.cern.ch/forum_thread.php?id=573 for its experiments with limiting the number of vCPUs in each VM.
19) Forums : Technical Support : Win10 Update Causes Problem? (Message 21143)
Posted 9 Aug 2016 by Jesse Viviano
Post:
Upgrading to VirtualBox 5.1.2 might break some VirtualBox applications. See https://github.com/BOINC/boinc/issues/1606 to see that a new wrapper could be needed for VirtualBox 5.1.2.
20) Forums : Technical Support : Virtualized tasks monopolize all cores in BOINC but only use around 40% of the CPU (Message 21098)
Posted 21 Jun 2016 by Jesse Viviano
Post:
I am sorry that I was unable to respond. I had to quit BOINC until the fall. My parents do not allow me to run it during the spring or summer because my computer running at full load will significantly heat up the room, and they only want the heat in the fall and winter. This includes GPU tasks.


Next 20