Steam miscalculates free disk space #205
Comments
|
How are you checking available disk space in this case? df? |
|
Nautilus and |
|
Since you're mentioning nautilus, I'll double-check that you ran df inside the container |
|
Where and how you mounting /var/home/data in the application? Just wondering since within the Flatpak $HOME/data is suppose to be a symlink |
|
The actual target directory is .data |
Bingo.
@alexlarsson, …? |
|
I would also strongly suggest adding some separate library folder instead of trying to mess around with data. Cloud Sync migration may delete your games from outside the container if you try to do anything to it. |
|
@nanonyme it is a separate library folder, just the LVM volume is named ‘data’. |
|
Ah, good. On another hand I just started wondering whether Flatpak should provide a way to protect folders against using --filesystem since it might break all amounts of expectations within applications.... @alexlarsson might be the best person to comment on that too |
|
So does that 3.9G map with some other disk's free space? |
|
@nanonyme not really, more importantly the 112K used doesn’t correspond with anything, I don’t know what that is. |
|
How much RAM do you have? |
|
Thought about that, but no; 8 gigs of RAM and (currently) 1 gig of swap. |
|
Might still be some space-limited temporary filesystem |
|
What do /etc/mtab and /etc/fstab say for you inside the container, btw? |
|
Neither of them are available, but this is the output of I mean what the hell is up with those mountpoints. |
|
You can't rely on /etc/mtab. Look at /proc/self/mounts (and maybe /proc/self/mountinfo) in the sandbox. |
|
Yeah, smelled heavily of tmpfs and maybe that part is fine. We're persisting . and turning data into symlink to .local/share which is also persisted. In this case filesystem is used to map a subdirectory to data though |
|
@jurf right? |
|
Also there's the question of where it pulls that 20GB used from |
|
@nanonyme I think you’re confusing XDG data and my ‘data’ partition. You can basically replace all instances of ‘data’ with ‘external HDD’. |
|
@alexlarsson, I can make very little sense of those files, sorry. Here they are: |
|
@jurf it's just that @alexlarsson still seemed to talk about XDG data |
|
OK, let’s start again. I have a partition by the name of On I tried changing the mountpoint to |
|
Yeah, I think the problem is that the steam flatpak doesn't expose the regular $HOME, as it uses --persist. So, what you see as /var/home/foo inside steam is in ~/.var/app/com.valvesoftware.Steam/ on the host. |
|
@alexlarsson it isn’t mounted in |
|
So, /var/home isn't your home, but /var/home/jurf. Ok, i misunderstood then. So, the issue is in the sandbox, this is the mount: I.e. the steamapps folder from the host is mounted on top of /var/home/data/games/steamapps. All the folders above that, such as /, /var, /var/home, /var/home/data, etc, are all part of the sandbox tmpfs root directory, and as such do not have the same free diskspace as the steamapps dir. |
|
@alexlarsson so if we have the exact location that was mounted with - - filesystem, should df show something sensible for that? |
|
@nanonyme yes, df will return the right values on /var/home/data/games/steamapps, just not on /var/home/data/games (or /var/home/data, or other parent). |
|
OK, so this is fixed by allowing Steam access to the directory that contains |
|
Upstream bug report: ValveSoftware/steam-for-linux#3881 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
It works fine with the home partition, but I have games on another one, which is mounted on
/var/home/data(this is Silverblue) and the games being on/var/home/data/games/steamapps. However, it looks like it’s mis-reading the partition size or something, because even though I freed ~10GB and extended the partition by another 50, it still only sees 3GB free.Wanted to check here first to see if it’s a cryptic bubblewrap issue before reporting it upstream.
I tried completely resetting the Steam installation, no help.
The text was updated successfully, but these errors were encountered: