Been running Tixati for a number of years, and I've recently been running into a problem, and I'm curious if anyone has seen something similar. My issue is that the speed of internal transfers on my network inside Tixati seems to crater after a while. Things like a file check, or file transfer from one folder to another drop to a snails pace. But if I close Tixati out all the way and reload it, all of a sudden everything is back to normal. Once whatever happens occurs, it stays that way till I exit and reopen. Last night I had it trying to transfer a torrent and at the rate it was going it would have been weeks. Closed and reopened and restarted the move and the transfer took around 5 minutes. I'm currently running around 2800 torrents (with about 50 active downloads), but even if I put them all offline except the ones actively downloading, it still happens. Ditto if I reduce the number of ones actively downloading. Obviously 2800 torrents is quite a load, so one thing I'm curious about is if anyone has seen any issues with very large torrent loads? I'm fortunate to have enough disk space that I pretty much seed darn near everything I download. Maybe it's time to thin the list out?
To give a full picture I'm running on a 64 bit Windows 10 VM installed on my Unraid server (8 cores & 32GB of ram assigned). I first noticed the issue on 2.81, so I tried downgrading back to 2.77 without any change. I have a dedicated hard drive outside of Unraid's array that I use as a scratch disk for the torrents. Then when they complete I have them transferring over to folders on the actual array. I am running PIA's desktop app (v2.5.1) and allowing local traffic on my LAN. All the folders involved are showing up as network shares to Windows.
Initially I was thinking that what I was seeing was a problem with my Unraid server itself, or with the unassigned devices plugin I'm mounting the scratch disk with. But in testing, even when Tixati is stalling I can run drive tests inside the VM and get normal full drive transfer speeds (same copying files into out of the same drive with Windows Explorer). But at the same time, I tried testing a file move from the scratch disk to another unassigned disk (keeping the array out of it), and was getting transfer times of like 1Mbs. Yet windows can copy to/from both drives and get rates more like 500Mbs. I closed Tixati out fully, reopened it and tried the same transfer, and I got the same 500Mbs transfer rate I was getting with normal file copies. But it seems like after being open for a long enough time, Tixati will get stuck with whatever is happening and at that point you can't realistically transfer files or check files until I close out and re-open. Torrents downloading seem to be unaffected, though I have noticed that if I check recent plots on the bandwidth page that overall downloads seem to start high (like 10Mbs) and eventually droop to 4-6MBs. That could easily simply be seeder availability though.
Anybody have any ideas? The fact that I can transfer files at normal speed with Windows Explorer while Tixati is stalling makes me lean towards it not being windows VM or server related, but I'm not ruling a windows or server problem out. Next thing on my list to try is loading a clean fresh VM, export my config from the existing VM to the new one, and see if the problem still happens or not. Will probably try that this weekend. Curious if anyone can think of anything else to try, or if anyone has ever run into anything similar.
