SSD space disappearing without being freed

[zEdS15B3GCwq]

Brief version:

The SSD I use as temp drive (-t) runs out of space all the time, despite having more than ample capacity for the number of processes using it. When I run chkdsk on it (just checking, NOT fixing), a lot of free space comes back. This isn't about TRIMming, though, because defrag /L (or defrag /O) doesn't have any effect. This may or may not have to do with Chia, but might be useful for others with a similar problem.

Details:

My setup:

• Windows 10 x64 + WSL2, 64GB RAM, 8/12 CPU cores
• node etc. running on Win
• 3 farmer processes in WSL2: `chia
• temp 1 (-t) drive is an NVMe SSD with 960 GB available
• temp 2 (-t) drive is another NVMe SSD with 450 GB available
• final destination is a spinning disk

Parameters I used for the farming processes:

chia plots create -t $CHIA_PLOTS_TEMP1 -2 $CHIA_PLOTS_TEMP2 -d $CHIA_PLOTS_FINAL -k 32 -n 10 -b 5120
chia plots create -t $CHIA_PLOTS_TEMP1 -2 $CHIA_PLOTS_TEMP2 -d $CHIA_PLOTS_FINAL -k 32 -n 10 -b 5500 -r 4

What' happening:

The available free space on the temp1 drive just goes down and down, 100+ GBs disappearing in an hour. Eventually, the farming processes run out of space. 3 farmers should not need more than 960 GB, especially when the -2 temp location is on another drive.

There are no leftover files; it's not like the recycling bin is keeping them. Even when all farming processes are stopped and their files deleted, the drive's total capacity seems to have shrunk. Somehow, the free space is not being restored. Interestingly, a simply chkdsk is enough to get the GBs back, but I have not enough knowledge to say why it works.

How to fix:

Regularly invoke chkdsk d: for the temp drive (assuming temp1 drive is d:) - no need /K switch

What doesn't work:

• defrag /O or defrag /L that does TRIMming on Windows
• killing the chia process and deleting the remaining files

Why I'm posting this

• Any ideas why this is happening?
• Ways to avoid this?