I'm monitoring an old laptop running as LAMP server. I have trouble understanding this VFS usage chart (sorry can't include images yet) from monitorix (the monitoring tool I'm using). I don't understand most of the terms I find on the internet, and I don't find a graph like this explained in easy words.

Is it correct that the yellow blocks (dentry) are free spots on the filesystem that can be used by VFS?

The blue blocks (inodes) are free spots on the filesystem that are ready to be used? Does the VFS make these because it expects writing operations?

Seems obvious but still: the pink blocks (files) are filled spots on the filesystem?

The VFS total usage was around 95-96%, goes up and down and ends at 100%. Should I interpret the total VFS usage as a cache that has dentries, inodes and files in it, and it's only a problem when the cache has too much files and not enough dentries to make into inodes?


If you look at the monitorix perl code in /usr/lib/monitorix/kern.pm we find it is getting these 3 statistics from the files:


These are described in man 5 proc. monitorix is calculating the following percentages:

  • dentry: % of allocated directory cache entries

  • file: % allocated file handles (number of files presently open). The maximum can be found from cat /proc/sys/fs/file-max.

  • inode: % of inodes allocated. When this gets over 100%, more inode space is created dynamically.

So you have a very low % of files open, an 80% full directory cache, and 100% inodes in use, but the latter will grow as needed. These values all concern the in-memory use of all filesystems, and do not apply to any given disc or partition. I dont think they signify any problem.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.