Just setup brand new UNRAID 6.7.2 two days ago on new machine, ran preclear on 2 4TB drives (successfully) and was adding all plugins (community applications, fix common problems, etc). I have only 2 shares (Movies and TV Shows) and just uploaded about a dozen movies to test before I load up the rest. On bottom left of "settings-> fix common problems" page it says this:

 

"Rootfs file is getting full (currently 83 % used) Possibly an application is storing excessive amount of data in /tmp. Seek assistance on the forums and post your diagnostics."

 

Well...here I am listening to instructions to not make a big mess myself before talking to everyone.

So what does it mean and what should I do?

Thank you.

deathstar-diagnostics-20190628-1634.zip

Greetings,

I think there's possibly two things going on. It appears you're running with 2G of RAM. While this is possible, because the root file system is running on that RAM it's complaining that it's getting full. With that little RAM, I'd think it might be somewhat expected to get that warning.

However, it is easy enough to check if this is high, but stable or high and getting worse. First thing I'd do would be to open the terminal and familiarize yourself with two commands - df and du.

I shall explain what they do and how to use them. To find out the filesystem sizes, issue df with an minus h (human readable) like this:

df -h

or if you just want to see the root filesystem that it's complaining about:

df -h /

You'll probably see this (from your logs):

Filesystem      Size  Used Avail Use% Mounted on
rootfs          812M  670M  142M  83% /

 

Now, if we want to figure out where the space is being used on / - we'll use du . This command essentially 'sums up' the disk usage in directories. I usually run it with the -h and often -s flags (for human readable and summarize respectively.) From the above df command, we already see that you're using 670M in your / (root) filesystem, so we could try to break that down a bit. Because some of the filesystems aren't relevant (you saw them with the straight df earlier) we'll exclude some via substitution as it's way easier and cleaner. You'll still need to ponder the output of this, and possibly drill down into directories to see if there's indeed an actual problem. So, to find out how the space is divided (there is a space after the } ) :

du -hs /* --exclude=/{proc,sys,dev,mnt,boot,run} /*

This should output a nice list with sizes for relevant root directories.

If indeed /tmp seems like it's the issue, you can check the size of things in it - in much the same way but the excludes won't be needed - since we're not checking the root file system.

du -hs /tmp/*

If you want to just see a specific directory size, say for /tmp we do it without the star:

du -hs /tmp

 

With these two commands I believe you should be able to get a handle on this situation and determine if it's stable but reporting high because you don't have much memory, or reporting high and getting worse.

Hope this helps,

Del

Edited June 28, 20197 yr by Delarius
fix inadvertent smiley