Hello everyone

I am manually monotoring partition sizes and available space from time to time.

I notized that on one of them the partition size have become smaller, and since I havent dont anything with the server, how can this be?

like from 59994 MB to 57213 MB

The config is a hardware raid controller with RAID 1 configuration.

Are you talking about the size of the partition itself or the disk space on the partition?
If disk space, a lot of MB can be used up in /tmp and also in /var/log
Your distro may be doing many other things behind your back as well that take up space...

Thank you for your reply

Yes it is the partition size it self that have become smaller
when i posted it was from 59994 MB to 57213 MB
as of today the number is 57215 MB

df -B M
prints
Filesystem 1M-blocks Used Available Use% Mounted on
/dev/sda5 11751 1013 10738 9% /
/dev/sda7 57215 30605 26610 54% /var
shmfs 505 0 505 0% /dev/shm

I suspect it is df that is giving different approximate values, as provided by the RAID controller.

When you create a file it will have a file size (number of bytes in the file), and a space on disk occupied. The amount of space taken will be equal or greater than the size of the file, dependant upon the block size.

For example with a block size of 512 bytes, that amount of space will always be taken up. E.g. for a 600 byte file, it will require two blocks and take up 1024 bytes on the disk.

I suspect that the RAID controller is returning a different number of blocks to take account of the partial used blocks. This would only be noticeable on a filesystem with a large number of small files, which may be the case for a var filesystem.

An alternative theory could be that the RAID controller is returning lower values as it is finding bad blocks on the disk. As this is a hardware RAID any bad block detection will be done outside of the OS, and so may not be reported by fsck. You would need to interrogate the hardware controller to see if there are bad blocks being detected, or it may be that these are logged in the syslog.

Thank you

The particulary system is running a cyrus pop3/imap server, so yes it is a lot of small files.
The size of the partition it self have been 100% stable on the same size for the last 6 months... only recently thease have changed to the current value.
I will check my syslog and consult my raid controllers documentation for how it handle disk errors

There is nothing in the syslog except a few write errors...
I cant access the hardware controller right now