I’m using Chia in the latest Docker container for 1.2.2 (ghcr.io/chia-network/chia:latest). I’m on Windows 10, and I have two 2TB NVMe drives for plotting. Previously I used Gentoo Linux to plot and encountered no issues with it, but I recently installed Windows 10 on another drive and decided to try plotting on Windows using a Linux container. However, I have only two plots running on a drive currently mounted in the Docker container, and running “du -sh /chiatmp0” (/chiatmp0 is the mount point of my Windows drive D:) shows the directory is only consuming 346GB. That sounds appropriate because on Gentoo I was able to plot at least 7 or 8 in parallel to a single drive. However, something is mysteriously using disk space because on Windows, the D:\ drive is reporting only 528GB of free space out of 2TB. Recall that the Linux container is reporting 346GB of used space as of this writing for the plots running on /chiatmp0. The math just doesn’t work out. There is nothing else on the drive, nothing in the recycle bin and no hidden files. It was newly formatted before I started plotting. I’ve trimmed it already. The space is not able to be reclaimed, and eventually, the drive runs out of space and plotting stops even though there are only two active plots, and I know how many plots this drive should contain in parallel. The only way I’ve found to reclaim that used space is to format the drive. Even after deleting the chia-generated temporary files from that directory, I don’t get back all of the free space from the drive. Formatting is the only thing that works. I don’t understand what’s happening.
Here is the output of “mount” in the Linux image:
D:\ on /chiatmp0 type 9p (rw,noatime,dirsync,aname=drvfs;path=D:;uid=0;gid=0;metadata;symlinkroot=/mnt/host,mmap,access=client,msize=65536,trans=fd,rfd=8,wfd=8)
E:\ on /chiatmp1 type 9p (rw,noatime,dirsync,aname=drvfs;path=E:;uid=0;gid=0;metadata;symlinkroot=/mnt/host,mmap,access=client,msize=65536,trans=fd,rfd=8,wfd=8)
Has anyone else seen this?