I wonder if there has been any thought of automating those tasks on a daily basis or other appropriate time frame.

But a large THANKS for getting the system back online!

Actually, WCG is handling what the client tells it in standard BOINC fashion (which assumes the host doesn't report a non-Windows O/S whilst asking for a Windows binary!), and in these cases the host is mis-reporting the O/S!

Unfortunately, the only ways to prove this are by being able to see what's in a scheduler request from one of these hosts or doing a code dive, though some of the examples I tried to include in my previous post would illustrate the issue (O/S information looking like Linux error reports!)

For an overview, I posted some hints as to what's happening on December 1st (including naming some BOINC source code modules of interest...). In the following discussions, TLD provided samples of various scheduler requests based on combinations on a WSL2 host -- those had correct information in each case.

The client needs to query the host O/S to get various details it includes in scheduler requests -- in some cases, a scheduler request is going in with platform details for Windows (proven by stderr reports for alleged Alpine Linux clients showing the MCM1 Windows executable!) but O/S details from whatever is associated with WSL (which might be an error message in some cases, often for tasks that fail in one way or another)...[*1]

Bottom line? It's not WCG's fault that standard BOINC trusts what a scheduler request tells it, or that standard BOINC doesn't check the application platform against the host platform to see it there's a major discrepancy. WCG (or BOINC maintainersl) will need to fix this if the client issues [with old wsl and/or Docker?] can't be resolved.

Cheers - Al

P.S. I tried to find some past posts that may have had suitable samples, but I couldn 't remember what thread(s) they were in and the Search function only seems to be able to find posts that were made after the brief outage and restart yesterday afteroon

*1 I don't think there's much point in "showing my work" on that topic; I have lots of examples, Forbidden and otherwise

I tested this with WSL enabled and disabled and will display here what are the relevant parts in scheduler_request:

WSL disabled:
....
<platform_name>windows_x86_64</platform_name>
<alt_platform>
<name>windows_intelx86</name>
</alt_platform>
....
<host_info>
....
<wsl>
<distro>
<distro_name>Ubuntu</distro_name>
<os_name>W</os_name>
<os_version>W</os_version>
<wsl_version>2</wsl_version>
<is_default/>
</distro>
</wsl>


WSL enabled:
....
<platform_name>windows_x86_64</platform_name>
<alt_platform>
<name>windows_intelx86</name>
</alt_platform>
....
<host_info>
....
<wsl>
<distro>
<distro_name>Ubuntu</distro_name>
<os_name>Ubuntu</os_name>
<os_version>Ubuntu 24.04.3 LTS</os_version>
<wsl_version>2</wsl_version>
<is_default/>
<libc_version>2.39</libc_version>
<docker_version>4.9.3</docker_version>
<docker_type>2</docker_type>
</distro>
</wsl>

Hi!
The time ( or the number of wus for the latest 24 hours) for the latest update on first page on Project Status is not updatting!?

https://stuxnode.com/images/wcg.jpg

Hans S.

The validation rate for tasks crunched before, during (cached tasks), as well as tasks crunched after the migration, is now dropping again. For me I can see that the number of "Pending validation" (crunched by both wingmen), is rising like a rocket.

So the result is of course, that my RAC (Recent Average Credit) is dropping like a stone.

Global statistics history:

[Statistics date	Total run time (y:d:h:m:s)	Points generated	Results returned
12/09/2025		97:347:12:26:03			275,079,998		461,822
12/08/2025		167:324:17:13:44		475,972,547		802,849
12/07/2025		202:086:01:36:11		570,838,318		962,729
12/06/2025		194:342:11:29:37		547,531,669		926,269
12/05/2025		192:023:13:57:44		536,398,856		908,430
12/04/2025		671:144:13:47:42		1,841,668,593		3,134,034