I'm trying to write an app_config.xml file so X number of 4-Gaia tasks and Y number of 6-Gaia tasks run at the same time so I can balance them out or overload one type versus another:

I started this from using a MilkyWay app_config file but it's not working:


<app_config>

<app_version>
<app_name>4_Gaia@home[20241019.1457], 4_Gaia@home[20241102.1102], 4_Gaia@home[20241130.0949]</app_name>


<avg_ncpus>15</avg_ncpus>

</app_version>

<app_version>
<app_name>6_Gaia@home[20241102.2234]</app_name>

<avg_ncpus>1</avg_ncpus>

</app_version>

<project_max_concurrent>3</project_max_concurrent>

</app_config>

I get the same error twice:
Mon 02 Dec 2024 10:48:16 AM EST | Gaia@home | Your app_config.xml file refers to an unknown application '6_Gaia@home'. Known applications: '4_Gaia@home[20241019.1457]', '5_Gaia@home[20241023.2232]', '6_Gaia@home[20241102.2234]', '4_Gaia@home[20241102.1102]', '4_Gaia@home[20241130.0949]'

I'm assuming I get the same error twice is because I have two app_version lines, can anyone help me? I have the numbers in the brackets because they are in the error message, I started without them but it didn't work then either.

Mon 02 Dec 2024 10:48:16 AM EST | Gaia@home | Your app_config.xml file refers to an unknown application '6_Gaia@home'. Known applications:

Looking at the error message it appears to have stopped reading the app name at the [.
Maybe need to put the name in quotes.

---

Paul.

This is what I used to currently control how many gaia_4 and gaia_6 tasks that I want to run since there are the only two apps currently available. Edit as needed when new apps appear. It works fine on my PC.

<app_config>
<app>
<name>6_Gaia@home[20241102.2234]</name>
<max_concurrent>4</max_concurrent>
</app>
<app>
<name>4_Gaia@home[20241130.0949]</name>
<max_concurrent>4</max_concurrent>
</app>
</app_config>

If you run out of Gaia 4 tasks then you'll only be running a single task. The client will see work available to run (it does not know about max_concurrent preferences) and not download any backup project at 0 priority. Just something to watch out for.

This is what I used to currently control how many gaia_4 and gaia_6 tasks that I want to run since there are the only two apps currently available. Edit as needed when new apps appear. It works fine on my PC.

<app_config>
<app>
<name>6_Gaia@home[20241102.2234]</name>
<max_concurrent>4</max_concurrent>
</app>
<app>
<name>4_Gaia@home[20241130.0949]</name>
<max_concurrent>4</max_concurrent>
</app>
</app_config>

Thank you VERY much!!!

If you run out of Gaia 4 tasks then you'll only be running a single task. The client will see work available to run (it does not know about max_concurrent preferences) and not download any backup project at 0 priority. Just something to watch out for.

I have setup a zero resource Project that would take over so it's okay, but thanks for the warning.

This is what I used to currently control how many gaia_4 and gaia_6 tasks that I want to run since there are the only two apps currently available. Edit as needed when new apps appear. It works fine on my PC.

<app_config>
<app>
<name>6_Gaia@home[20241102.2234]</name>
<max_concurrent>4</max_concurrent>
</app>
<app>
<name>4_Gaia@home[20241130.0949]</name>
<max_concurrent>4</max_concurrent>
</app>
</app_config>

Thank you VERY much!!!

It doesn't work though, it STILL comes up with the same error message

Mayby \[ and \] ....

You probably have already done this but no harm in confirming if you already did read the updated app_config.xml file via the Options → Read config file in the Boinc Manager menu?

The app name is what is shown in my client_state.xml file. See below for mine. Seems to be correct app names to me.
<app>
<name>6_Gaia@home[20241102.2234]</name>
<user_friendly_name>6_Gaia@home</user_friendly_name>
<non_cpu_intensive>0</non_cpu_intensive>
</app>

<app>
<name>4_Gaia@home[20241130.0949]</name>
<user_friendly_name>4_Gaia@home</user_friendly_name>
<non_cpu_intensive>0</non_cpu_intensive>
</app>

Hey mikey, you did remember to stop BOINC do your app_info and then restart BOINC. You may of overlooked this step perhaps

Conan

---

Conan, this is an app_config. An app_info needs a client restart.

If you run out of Gaia 4 tasks then you'll only be running a single task. The client will see work available to run (it does not know about max_concurrent preferences) and not download any backup project at 0 priority. Just something to watch out for.

I have setup a zero resource Project that would take over so it's okay, but thanks for the warning.

It won't since the client still shows that you have work available from a non-zero project and will not download work.

Known applications:
4_Gaia@home[20241019.1457]
5_Gaia@home[20241023.2232]
6_Gaia@home[20241102.2234]
4_Gaia@home[20241102.1102]
4_Gaia@home[20241130.0949] ----

I'm guessing only the last one is the correct one for Gaia 4 as it matches the date on the apps.php page of 11/30. If not, then keep adding apps to the config file for any "known" apps the log shows.

Conan, this is an app_config. An app_info needs a client restart.

If you run out of Gaia 4 tasks then you'll only be running a single task. The client will see work available to run (it does not know about max_concurrent preferences) and not download any backup project at 0 priority. Just something to watch out for.

I have setup a zero resource Project that would take over so it's okay, but thanks for the warning.

It won't since the client still shows that you have work available from a non-zero project and will not download work.

.

Simple solution.
Run this project in a VM or multi BOINC install.
No BOINC queue management headaches.
The queue was setup for a single core machine and does not allow multiple queues assigned to groupings of cores. It's clunky and needs to be completely reworked.

Conan, this is an app_config. An app_info needs a client restart.

[quote]If you run out of Gaia 4 tasks then you'll only be running a single task. The client will see work available to run (it does not know about max_concurrent preferences) and not download any backup project at 0 priority. Just something to watch out for.

I have setup a zero resource Project that would take over so it's okay, but thanks for the warning.

It won't since the client still shows that you have work available from a non-zero project and will not download work. [/gquote]

It changes on a daily basis so I'm not worried about it or finding a backup project

Known applications:
4_Gaia@home[20241019.1457]
5_Gaia@home[20241023.2232]
6_Gaia@home[20241102.2234]
4_Gaia@home[20241102.1102]
4_Gaia@home[20241130.0949] ----

I'm guessing only the last one is the correct one for Gaia 4 as it matches the date on the apps.php page of 11/30. If not, then keep adding apps to the config file for any "known" apps the log shows.

Why did they name it like that, why not just 6_Gaia@home and adjust the tasks as needed that's what every other project does.

Conan, this is an app_config. An app_info needs a client restart.

If you run out of Gaia 4 tasks then you'll only be running a single task. The client will see work available to run (it does not know about max_concurrent preferences) and not download any backup project at 0 priority. Just something to watch out for.

I have setup a zero resource Project that would take over so it's okay, but thanks for the warning.

It won't since the client still shows that you have work available from a non-zero project and will not download work.

.

Simple solution.
Run this project in a VM or multi BOINC install.
No BOINC queue management headaches.
The queue was setup for a single core machine and does not allow multiple queues assigned to groupings of cores. It's clunky and needs to be completely reworked.

I guess I could but was hoping a simple app_config would solve the whole thing!!

Simple solution.
Run this project in a VM or multi BOINC install.
No BOINC queue management headaches.
The queue was setup for a single core machine and does not allow multiple queues assigned to groupings of cores. It's clunky and needs to be completely reworked.

I guess I could but was hoping a simple app_config would solve the whole thing!!

That server is running a custom stripped down Windows 10, with minimal services and maximal RAM available, so the subsystem fails.
I just built a minimal AntiX Linux VM and maintain it (good learning experience for low skill Linux tech like me)..
That VM has been a nice addition as it sits there waiting for a Linux only project to generate rare WU's and not a serious issue to share 16 cores of the 48 available, for a while, unless it's prime number searching. Those are a race.

Oh, some projects ignore the 0 resource anyway and might send you tasks. NFS is one. Yoyo used to.

(Nice seeing you here. We WUPropers get around!!!).

Simple solution.
Run this project in a VM or multi BOINC install.
No BOINC queue management headaches.
The queue was setup for a single core machine and does not allow multiple queues assigned to groupings of cores. It's clunky and needs to be completely reworked.

I guess I could but was hoping a simple app_config would solve the whole thing!!

That server is running a custom stripped down Windows 10, with minimal services and maximal RAM available, so the subsystem fails.
I just built a minimal AntiX Linux VM and maintain it (good learning experience for low skill Linux tech like me)..
That VM has been a nice addition as it sits there waiting for a Linux only project to generate rare WU's and not a serious issue to share 16 cores of the 48 available, for a while, unless it's prime number searching. Those are a race.

Oh, some projects ignore the 0 resource anyway and might send you tasks. NFS is one. Yoyo used to.

(Nice seeing you here. We WUPropers get around!!!).

SRBase does Prime Numbers too and has tasks available! But NOT for Linux only, they have a Windows app as well.

The app_name change daily? They match the app version date/time that's still listed on apps page.
6_Gaia@home[20241102.2234]
2 Nov 2024, 21:34:31 UTC
4_Gaia@home[20241130.0949]
30 Nov 2024, 8:49:28 UTC

I'd rather install Linux via WSL vs a VM.

Several projects will send a task or couple at zero weight like NFS. NF is another. They only do it when there is no non-zero work from what I've seen.

I'd rather install Linux via WSL vs a VM.

Just installed Puppy Linux to a 128GB thumbdrive and will boot it, install BOINC and attach Gaia/WUProp, then see if it will successfully run Gaia 4 WU's on another server (still confused why AntiX runs about 10% of those #4 WU's instead of 100% failures).

Gaia has been very intermittent, so if the VM detects work, then I could use the thumbdrive to bring more cores to the project for a week or so.