Solaris

Hi

Can anyone tell me the details (or) purpose of the minfree(10%) space in disk.


Thanks in advance
MaroV

When you do `newfs`, there's the -m switch, with which you specify the
procentage of the disk which will remain reserved ("minfree"). When the
disk fills up to this threshold, any nonprivileged user that tries to write
to the FS will be denied, getting a message that the disk is full. The
reality of it is that the disk still has free space, but only root and
processes can use that space to write to.

Hi Incrdible

Can you tell me what are the processes reside in min free space.

Thanks
MaroV

minfree is the percentage of a file system reserved for the superuser or privileged processes. The value is 10% by default on UFS file systems.

/usr/sbin/fstyp -v raw_device | grep minfree

ex.
# /usr/sbin/fstyp -v /dev/rdsk/c0t0d0s0 | grep minfree
minfree 10% maxbpg 2048 optim time

vr_mari,
minfree may turn out useful, when your disk becomes full up to last byte. In that case you'd want to log in to system to kill/delete what caused the problem. But logging in to system triggers additional entries to utmpx, wtmpx and perhaps messages -> which require space. It's good to have some extra spare space in case something like the above hits your system.
I can't recall it right know for sure - but besides what Incredible described minfree may have something to do with managing fs fragmentation.

yes, mine was on managing fs fragmentation

minfree's purpose is to minimize fragmentation. If 10% of the the disk is free, then either all of the cylinder groups have about 10% free or some cylinder groups have a lot of free space. Under either condition the the original Mckusick filesystem code would allocate files with little fragmentation. The idea is that fast performance is worth a little disk space. 10% is probably excessive with today's large disks, but disk space keeps getting cheaper and I have never bothered to experiment with smaller values.

Allowing root to continue to use the disk past minfree is a bonus. But bear in mind that if root truely fills the disk, the last files written will almost certainly be very badly fragmented.