D-did I win? [the Old M1 8gb SSD Swap problem]

[Kaikidan]

Never had problems before, today was listening to music on PlexAMP using SoundSource for an hour or so and for the first time the system has run out of memory, I opened the activity monitor to check and oh my, somehow there was two processes using 16 and 25gb of memory... on a 8gb machine and too much for even a 16gb machine, I strongly suspect that SoundSource is leaking memory somehow, since RAM was not a problem until now, and after installing the app now it spikes abnormally high. Second picture is how it normally behaves on my daily tasks.

 

[jav6454]

I think you won big. 25GB on a single process that is not a VM, that is crazy.

 

[mi7chy]

I've experienced memory leak using over 13GB on 16GB MBA M1 with both Monterey and Big Sur and no apps running so highly suspect it's MacOS. Was hoping rolling back from Monterey to Big Sur would fix it but nope.

 

[steve217]

Never had problems before, today was listening to music on PlexAMP using SoundSource for an hour or so and for the first time the system has run out of memory, I opened the activity monitor to check and oh my, somehow there was two processes using 16 and 25gb of memory... on a 8gb machine and too much for even a 16gb machine, I strongly suspect that SoundSource is leaking memory somehow, since RAM was not a problem until now, and after installing the app now it spikes abnormally high. Second picture is how it normally behaves on my daily tasks.

It used 25GB of swap and the pressure was only yellow?

How did it function during the leak? Did you get a message and then check Activity Monitor?

 

[Kaikidan]

It used 25GB of swap and the pressure was only yellow?

How did it function during the leak? Did you get a message and then check Activity Monitor?

strangely enough it was acting fairly normal, until the out of memory pop up appeared, then audio lagged pretty bad, browser force-closed up, but still pretty responsive even managed to open the activity monitor no problem. I found really weird too that the memory pressure was on the yellow with so much swap, which leads me to conclude that it was some sort of memory leak on the audio process and not a case of really using up ram to the max on aplications.

 

[jav6454]

strangely enough it was acting fairly normal, until the out of memory pop up appeared, then audio lagged pretty bad, browser force-closed up, but still pretty responsive even managed to open the activity monitor no problem. I found really weird too that the memory pressure was on the yellow with so much swap, which leads me to conclude that it was some sort of memory leak on the audio process and not a case of really using up ram to the max on aplications.

Have you been able to replicate the issue?

 

[Kaikidan]

Have you been able to replicate the issue?

I will try tomorrow on another listening session, the two possible culprits are plexamp and SoundSource, taking into account that system behaved normally until before I installed SoundSource and plugged my external AMP/DAC, I leads me to believe it has to do with SoudSource since it installs a kernel extension that needs permission from the recovery boot menu.

 

[jav6454]

Keep us posted then.

 

[satcomer]

Did you check the privacy System preference to allow the Hack install?

 

[jdb8167]

I had “systemstats” taking up 28 GB yesterday. Had to reboot to reclaim the memory. Logging out and back in didn’t work and the process was owned by root so I didn’t want to just force kill it.

I also showed yellow memory pressure with 8 GB of compressed memory and 19+ GB of swap. This on a 16 GB M1 MacBook Air. Performance was fine throughout.

 

[Kaikidan]

Update: Today memory leak started to happen again, it was at almost 1,5gb on coreaudiod. I force closed the process and it started building up again. then I force closed SoundSource and the process, now it's sitting at 20mb. So definitely the app is leaking memory, I've send a bug report to the company.

EDIT: After contacting the support, they offered a development build for 5.5.4 which seens to fix the problem. For now it's behaving fine.