Hi,
i run two Android devices with WGC for several weeks without problems.
Now i switch machines to Einstein and all WUProp units crash at around 25% progres.

106 WUProp@Home [error] Couldn't destroy shared memory: shmctl() failed
105 WUProp@Home [error] Couldn't detach shared memory: shmget() failed

<core_client_version>8.2.11</core_client_version>
<![CDATA[
<stderr_txt>
 initialisation
attach_shmem: not supported on this platform: Success
 Can't set up shared mem: -1. Will run in standalone mode.
lspci: /sys/bus/pci/devices: Permission denied
Erreur assignation taille wu
Erreur assignation taille application
 Aucun projet actif
 Aucun projet actif
.
.
.
 Aucun projet actif
 Aucun projet actif
 initialisation
attach_shmem: not supported on this platform: Success
 Can't set up shared mem: -1. Will run in standalone mode.
 called boinc_finish(0)

</stderr_txt>

---

Greetz
Chris

I see the same error message in all my successfully completed Android tasks.

I had posted in number crunching about something similar.
https://wuprop.statseb.fr/forum_thread.php?id=753&postid=12356

The English translation of the error message my apps keep showing in the std-err message that's reported on the website.:
WU size assignment error
Application size assignment error

The problem is that no project admin tells something about.
The units with the two french erreurs will be counted normal on one of my devices (Boinc 8.2.8).
But now i have two devices where units crash and wil not count.
One of them don't get any new tasks at all from WUProp.
Both machines have Boinc 8.2.11 may that is the problem ?

The problem is that no project admin tells something about.
The units with the two french erreurs will be counted normal on one of my devices (Boinc 8.2.8).
But now i have two devices where units crash and wil not count.
One of them don't get any new tasks at all from WUProp.
Both machines have Boinc 8.2.11 may that is the problem ?

I agree with your statement "The problem is that no project admin tells something about." I'm not sure if Sebastian or any other possible admin reads the forums.

I don't think it's something specific to version 8.2.11 is the problem. If I recall I started noticing my error messages sometime after installing version 8.0.x. I currently run 8.2.12 on my 3 Android phones.

Tried beta 8.2.15 but had to uninstall and downgrade to .12 again due to BOINC Android would only run one or two tasks before receiving the "swap file error message,"

On the 2 devices you are having problems with make sure the box is checked for "Enable Remote Control" Under the Advanced options in the Android app. If not checked the WuProp tasks will error out and fail.

I see your device #351582 running Android
4.14.116 (Android 10)
is error out all tasks with
"Aucun projet actif"
Translated to 'No active projects".
If no other project tasks other than WuProp are running or the remote control is not enabled you won't get any credit for WuProp.

all my devices are remote controlled because i use Boinc tasks...
One of the two devices runs normal again now, since Einstein apps show right estimated runtime...
I run a benchmark in Boinc and after that i saw very wrong estimated run times.. but now its over.
The other machine is a cell phone and have 4 GB RAM only, maybe it is problem there...
106 WUProp@Home [error] Couldn't destroy shared memory: shmctl() failed
105 WUProp@Home [error] Couldn't detach shared memory: shmget() failed
This device don't get task.. every update request is answered with

568 WUProp@Home Not accepting requests from this host
569 WUProp@Home Project requested delay of 86400 seconds

4 GB is more than enough RAM to run BOINC for any of the current projects that support Android. None of mine have more than 4GB.

Your Android device I mentioned in the previous post #351582 was working up until 16 Jul 2026, 4:12:08 UTC. Every task since then error out with only 2 seconds run time. Looking at the task list for the device... Something changed after that time.

Maybe try powering down (hard shut down - power off) and restarting the device if that hasn't been done in a while. I usually power down and restart my 3 Android phones at least once a month to keep them running smoothly.

How many tasks are you trying to run st the same time?

For my two 8 core phones, I can only run 3 consecutive tasks + one WuProp before they go to thermal overload.

Your problem device(s) won't get new task for 24 hours from the last invalid one returned and will keep on the 86,400 second time out each time until it returns a good/valid tasks.

4 GB is more than enough RAM to run BOINC for any of the current projects that support Android. None of mine have more than 4GB.

Your Android device I mentioned in the previous post #351582 was working up until 16 Jul 2026, 4:12:08 UTC. Every task since then error out with only 2 seconds run time. Looking at the task list for the device... Something changed after that time.

Maybe try powering down (hard shut down - power off) and restarting the device if that hasn't been done in a while. I usually power down and restart my 3 Android phones at least once a month to keep them running smoothly.

How many tasks are you trying to run st the same time?

For my two 8 core phones, I can only run 3 consecutive tasks + one WuProp before they go to thermal overload.

Your problem device(s) won't get new task for 24 hours from the last invalid one returned and will keep on the 86,400 second time out each time until it returns a good/valid tasks.

Yes, something changed... WCG server sucks and machine was empty. Then i add Einstein and the problems begin.
I run 3 units + WUP at once.
After the first 2-3 crashed units (WUProb, Einstein run fine) i restarted the device indeed.
Today i will try again if i get a task.... ok i see there is still one task in progress aldus server. But this task i delete at the device, i cpl. reset WUProp. Maybe therefore i don't get any new task.