Rclone cmount silently stops

[ncw (Nick Craig-Wood)]

@Floris

Test 1: opening mounted remote from launchagent mount

@Floris

Wakeups:          45001 wakeups over the last 7 seconds (6545 wakeups per second average), exceeding limit of 150 wakeups per second over 300 seconds

My understanding (which may be wrong) is that the kernel kills the process at this point.

@Floris

Test 2: terminal mount with rclone log -vv

@Floris

Apparently, whatever is the reason for the mount to disappear is not related to the amount of files in the folder, as I first thought. The prepared folder with many different types of files worked fine. Instead, the mount appears to have crashed through sustained requests from my photo app test.

That agrees with my understanding. Rclone gets killed off when it reaches that wakeups/second limit. However wakeups / s are entirely due to the users of the mount not rclone itself.

I can't see anything obvious in the logs.

A

@Floris

I will attempt to comb through the huge logfile later with some terminal commands. It might help if you have some suggestions what to look for @ncw if that is not a bother.

What I'd really like to know is why did rclone die. This should be at the end of the log file.

I made a test binary with a new flag --mount-tpslimit - can you try it with --mount-tpslimit 100 and see if that works. If it doesn't then halve it until it does. This might slow things down a lot but hopefully it will stop rclone waking up too much.

Ideally we want to opt rclone out of the wakeup protection, but I don't know how to do that - any ideas gratefully received!

v1.55.0-beta.5156.cbe049b52.fix-macos-mounttps on branch fix-macos-mounttps (uploaded in 15-30 mins)