Some misconfiguration might can lead to file record segment corruption.

[sh20000sh]

Recently I had couple of fatal filesystem failures. It might be just because system running yt2009 is on old hardware, but I had synthetic case that occurred by yt2009 on earlier days - which was changing name of video file to different resolution and playing it with flash player.

I'm not sure about normal use can make this happening but just as note.

[ftde0]

you're not using the same old hard drive that keeps failing.. right? yt2009 does read and write a bit, especially if you're downloading, but that's what you have to do to support old setups

[sh20000sh]

Frankly speaking, looks it's less likely hardware failure. Anyway, it can be fixed with chkdsk /r when running before things become too complicate.

[ftde0]

please try with a different hard drive before coming to such a conclusion. i'm having no problem hosting yt2009 on an 18 year old computer with near-everyday usage with a swapped hard drive (the old one straight up showed errors on startup to back up data immediately)

[sh20000sh]

please try with a different hard drive before coming to such a conclusion.

I understand. But I recently had actual hard drive failure on other system and case I experienced on yt2009 running had different behavior with that. It was happened like once in two months so nothing is certain.

If these things are real, I think cause might be conflict with auto-maintain, such like #310 .

[ftde0]

the error shown in that issue was related to the downloader trying to remove a temp asset that was already gone. if that was caused by auto_maintain

• you'd have to be VERY unlucky (after the initial run on instance start, auto_maintain does its thing once every 4 hours as long as the frontend is running)
• you should have an explicit notification in console logs that auto_maintain removed files and freed up space

or something else (the file being downloaded for something else, manually deleted etc).

hard drive failures are wild. i've seen a few, there were cases where after reinstalling windows it would work fine for a week and start showing bluescreens afterwards (that behavior was consistent), other times i've seen drives become unusably slow close to their failure point, and that one time i saw one where new files would straight up get corrupted or blanked out.

in any case there are no other reports of yt2009 specifically causing filesystem failures, but since this is a discussion this won't hurt to keep open. if nobody else reports anything else in a span of a few months, then that would mean the issue is somewhere on your end

[sh20000sh]

you'd have to be VERY unlucky (after the initial run on instance start, auto_maintain does its thing once every 4 hours as long as the frontend is running)

Error itself was caused by that I know. I sometimes watch same video few hours later.