In reply to Dave's message of 16 Jul 2025:

I seem to have that problem with WCG tasks on 8.3.0
Currently no tasks available but I have been running ARP tasks for months and estimated times are still over double actual run times.

Edit: If confirmed when WCG tasks start appearing again, I will flag it at git-hub. (Assuming it isn't already there. I haven't checked yet.)

Just to confirm this is a Bug with BOINC 8.2.4
Downgraded to v8.0.2 on one system, Remaining (estimated) GPU times are moving around again as work is returned and new work downloaded.
The times are still frozen on other system running v8.2.4

---

Grant
Darwin NT.

Bug report filed.

I did look and didn't find an already filed one.

In reply to Dave's message of 19 Jul 2025:

Bug report filed.

Thanks for that.

---

Grant
Darwin NT.

Another possible bug posted over at Rosetta.

Win 11, WSL has multiple distros.
Picks "last" instead of "default"

Bug:
Operating System denoted in "Your computers" table is displayed as the last alphabetical GUID of the WSL distro instead of default GUID at HKEY_CURRENT_USERSOFTWAREMicrosoftWindowsCurrentVersionLxss.

Expected:
Default distro is chosen/displayed.

Workaround:
- Find the GUID in the registry of the distro you want Rosetta to use.
- Rename GUID to sort last alphabetically (e.g. change to ffff... prefix).
- Set value of HKEY_CURRENT_USERSOFTWAREMicrosoftWindowsCurrentVersionLxssDefaultDistribution to same new guid.

Value of Operating system will now be correct.

---

Grant
Darwin NT.

I have multiple projects, some may use VB (have in the past).

Is anything actually running today, no. But may be in the future.

So, wait and see or abandon them and wait, until when?

Richard

The when part of your question is very much in the hands of the individual project, so you will have either ask them the question or monotor their boards to see what are doing and planning....

The only BOINC project I found that uses Docker is BOINC Central and, as far as I can tell, the Docker part, is just for testing right now.

Thanks, that is helpful.

In reply to Grant (SSSF)'s message of 19 Jul 2025:

In reply to Dave's message of 19 Jul 2025:

Bug report filed.

Thanks for that.

davidpanderson
left a comment
(BOINC/boinc#6449)

Nothing changed from 8.0.2 to 8.3 in this area, and as far as I know the client never has changed runtime estimates based on times of completed jobs

I have replied to the list that that is not my experience. Might be good if one or two other people confirmed what has changed over on git-hub. (If you don't have an account it means setting one up. You only get emails from items you have contributed to so you won't get spammed by email on any other issues being worked on for BOINC.

Might be good if one or two other people confirmed what has changed over on git-hub.

Done.

---

Grant
Darwin NT.

I've recently built a new Windows 11 machine and installed the 8.2.4 BOINC Manager. This has been a problem filled build, so there were several iterations to get to the present point.

Initially I installed an older client and BOINC seemed to run correctly, even on the integrated GPU of the AMD7950X. I then noticed there was a new version and installed BOINC 8.2.4. I received the message in a notice that I should install Docker. Since this was a new build, I decided to go ahead and do the PODMAN install. Note that BOINC was installed and running before the PODMAN install. After completing the build with an RTX 4070 GPU, the tasks that had been downloaded were completed, but no new tasks have been received.

I have two older machines that continue to get new tasks. They have been updated to 8.2.4, but PODMAN has not been installed.

I'm now thinking that the projects I run, DENIS@home, Einstein@home, gpugrid, Rosetta@home, SiDOck@home and World Community Grid don't support Docker/PODMAN and don't have any tasks that will run on a system set up for this.

I'm going to try two things.

First, I'll delete all of the projects from this machine and then add them back. I hope this will reset the delivery request in some way and I'll get new tasks.

Second, I'll delete all the PODMAN stuff and reload an older version of BOINC with virtual box. After the older version of BOINC is running, hopefully, I'll upgrade to 8.2.4 but without PODMAN.

I'll update this after I make the changes.

---

Currently very few projects need virtualisation, and even fewer will run virtualised under docker/podman (lhc and boinc central are the only two yhat I can think of).

In reply to clcarter1999's message of 24 Jul 2025:

I'm going to try two things.

First, I'll delete all of the projects from this machine and then add them back. I hope this will reset the delivery request in some way and I'll get new tasks.

Second, I'll delete all the PODMAN stuff and reload an older version of BOINC with virtual box. After the older version of BOINC is running, hopefully, I'll upgrade to 8.2.4 but without PODMAN.

It's probably too late- but don't do that.
Installing or not installing Podman has nothing to do with work fetch.

All that's needed at this stage would be to look at the Event log messages to see what's going on there- such as local settings (eg No new Tasks, network activity suspended, suspend if in use settings etc) overriding web-based ones.


EDIT-
FYI- DENIS@home presently has no work
Rosetta presently has no work.
SIDock has no work.
GPU Grid has no work
WCG- could be anything from servers down, network down, to has no work, to who knows what- you need to check there to see what is going on at any given time, day by day (hour by hour, minute by minute...).
Einstein appears to have work and the servers & site are alive (at least according to it's status page)- so you should be able to det some work from them, if what's available is available for your OS.

---

Grant
Darwin NT.

I got it working. I read through the responses I received and I want to thank you for those.

Deleting, rebooting and re-adding projects had no effect.

I uninstalled BOINC and PODMAN, rebooted, then reinstalled an old version of BOINC. Again no tasks received overnight.

I deleted BOINC again and scrubbed the registry of any reference to BOINC. Don't do this if you are not familiar with how these registry entries could affect your computer. It could be ugly.

I reinstalled the older version of BOINC. I added the Einstein project and updated. I quickly received several tasks. I then added BAM and synced to get all of my projects.

I now have 28 Einstein tasks running. My other machines are running mostly Einstein tasks with occasional Rosetta tasks coming in slugs.

At some point I'll update to the newer 8.2.4 BOINC that my other machines are running.

Thanks again.

Hi, I updated Windows 11 (to July 22, 2025—KB5062660 (OS Build 26100.4770) Preview) and shortly after that installed Boinc 8.2.4 and Podman. I'm noticing now that my screen saver does not come on, no matter which one I set it to. The control panel dialog box behaves oddly, changing the screen saver setting just with mouse movement without choosing anything on the drop down.

Two issues I have seen with 8.2.4 and Docker.

1 - Parallel communication with projects if one project is down doesn't seem to exist. At this very moment ODLK2025 is down and tasks are piling up trying to upload. I am also running some smaller Yafu tasks plus WUProp. Due to ODLK being down, the Yafu and WUProp tasks will not report. ODLK is stuck on Work fetch according to BOINC Tasks when viewing the projects tab and Yafu/WUProp are updating but never do. I reset ODLK project and dumped all my ODLK tasks. Yafu and WUprop immediately updated and I received new tasks. In the past few weeks with Goofy (an NCI project) and a prior instance with ODLK2025 have caused the exact same issue. I don't recall seeing this in years past when projects go down but communication between all projects and client should not be held up if a single project goes down.

2 - Docker memory reported to the client is completely bogus. I switched from another project with tasks using 10-15mb to BOINC Central and tasks were 300-400mb and the memory usage in htop did not change. Yesterday, CDPN released some work (1.3 to 1.4GB memory used per task) and I downloaded some tasks on a couple PCs that were already running BOINC Central. Those Docker tasks were reported as using 500-600mb but once the CDPN tasks started the memory reported per docker task started to climb rapidly. Tasks started to suspend due to what the client thought was a lack of memory. I suspended all of the docker tasks and let a single docker task continue with htop running. The docker task shot up to over 25.5 GB (Yes giga) in according to the client all while htop never saw an increase and only reported 20GB in total used while running multiple of those CDPN tasks. I tried it again on a 2nd PC and I saw over 38GB of reported memory usage prior to suspending it. Why does the reported memory usage of docker depend on what other tasks are running?
https://boinc.berkeley.edu/central/forum_thread.php?id=34

The docker task shot up to over 25.5 GB

I shudder to think what it will report when the multi-core CPDN tasks arrive which really do use over 25GB of RAM!

My windows PC on 8.0.2 got up over 100 pending upload tasks then it failed to upload a Yafu task. Reset ODLK and the yafu uploaded/downloaded another.

God evening; I managed to install 8.2.4 + docker/PODMAN on my MacBook Pro.

To install the windows version of bionic W10 and 11 is far too complex for me.

Right now I plan only to install BOINC 8.2.4 without Docker/PODMAN. What problem will this lead to?

Kenneth in Stockholm.

Right now I plan only to install BOINC 8.2.4 without Docker/PODMAN. What problem will this lead to?

None. The only work that uses Docker/PODMAN is experimental to iron out problems with it. The long term plan for it is to have Docker installed by default by BOINC enabling projects that are picky about the software environment to run everything in a VM which will have exactly the right version of Linux but stripped down without superfluous bits to it. There are no REAL projects using it and probably won't be for a few years at least.