Problem : Rclone mount crash and die after a few days. Pretty light usage. I'm seeding from the mount, but I limit the speed, I don't use full cache, nothing hardcore... My computer is not stressed when the crash happens....... And it's only since a few versions.
Two crashes in the log, one a few days ago, and one a few minutes ago :
--fast-list does nothing on a mount so you can remove it.
but add it to rc vfs/refresh recursive=true --fast-list
--use-mmap
given the documentation does not work well on all platforms usually means windows os.
and if you read the logs, there seems to be memory errors runtime.sigpanic() runtime.memmove(0x23600620000, 0x23606150000, 0x1000)
i suggest to remove that flag and test.
About nmap, I'll try without it. But I used it for months until this point, without problems.
I've 32gb of ram. I never paid attention about winfsp memory usage, and rclone varies between about 100mo to 500mo, depending on the the number of transferts at the same time. In any case my ram is far from full.
Well, that's why I posted this with the log, so it can help finding what's wrong... I'm not complaining or anything... And I participated in the new vfs cache mode topic, I know the drill...
Like @asdffdsa said, start with a much simpler mount. I've been using the latest beta of both rclone and WinFsp for close to four years now. Never had a mount crash. I'm not saying it can't happen, but it very likely has something to do with some of your flags.
Oh, and --fast-list doesn't do anything for the refresh command either.
I don't have a use case for cache or writing to the mount, so mine is as follows:
Well, my mindset is, the crashes are pretty new, and I've used theses flags for a long times without a problem. If a flag / settings is causing a problem now, it's still usefull to report it and try to fix it, if the problem is located
It's definitely useful to report crashes without a doubt, I think everyone agrees with that. Thanks for reporting it!
On a separate note though, removing none useful flags just makes things cleaner. I only add a flag if I have a reason to add it and otherwise, I use defaults as my 'simple' approach to rclone and pretty much everything.
The day of the release, so 22 days ago. I don't believe it's linked when I'm look at what changed vs the non beta, but I'll revert to stable if I've crashes again (right now i'm just testing without nmap, i prefer testing one thing at a time).
Sadly on Windows, I have no idea where you'd find that as depending on how you started it, that is what would capture it as on Linux, it would be systemd.
Do you see the stack trace in the log you created?