I am seeing call traces in my system log related to page allocation stalls/failures. The system seems to be running just fine, but the errors happen repeatedly.

 

Description:

I see call traces in the logs showing page allocation stalls in plex, btrfs, and other processes on a semi-frequent basis. I am not sure of the cause or how to reproduce. 

 

Other information:

Please note you can ignore the Plex segfaults, they are an apparent known issue with Plex and will be fixed in the next public release.

 

Diagnostics attached.

 

P.S.:

The diagnostics plugin leaks information due to the mover application. The mover leaks filenames into the system log. The mover should likely either have its logging changed or the diagnostics packager should anonymize the filenames when packaging.

tower-diagnostics-20170403-1936.zip

Edited April 4, 20179 yr by ssb201
clean up formatting

You can turn off the mover logging under Settings->Scheduler.

If you have a lot of memory pressure you can try installing the Swap file plugin listed in Community Apps.

Thx @itimpi. I completely missed that setting, though I would like to be able to log the mover running, just not the filenames moved. I will have to see if the start and stop is still logged when I turn it off.

 

@limetech I will give it a shot, but I am not sure why I would be experiencing any memory pressure. I currently have a single VM running with dynamic memory up to a limit of 24GB of RAM on a 64GB box. When I look at Dynamix Stats I see the vast majority (46+GB of RAM is devoted to cache).

24 minutes ago, ssb201 said:

I currently have a single VM running with dynamic memory up to a limit of 24GB of RAM on a 64GB box

 

You could try setting the memory to a static min/max (ie: 8gb min/8gb max). Not sure if it will help, but I believe dynamic memory settings can lead to some instability issues...

Thanks, I will give that a shot as well.