Once upon a time, kernels and filesystems had a few more bugs than
presently (hence more crashes), disks weren't as reliable (hence more
data loss), and they were slower (fsck took forever). That's where
the "old habit" came from. Many times, each of those filesystems
lived
on its own (256 MB!) disk pack.
While these factors are less important lately, they still exist. I've
recently
worked with critical file servers that would take 20 minutes to fsck,
and if
they had a simple, single filesystem, they would take hours.
These arguments are relevant for the "MD/LVM" vs "lots of filesystems"
debate. In the multi-disk,single filesystem configuration (w/o RAID)
the
loss of a single disk (due to hardware failure, for instance) will
cause all
your data to be lost, and you must restore it all from backup.
It boils down to an admin's decision to trade off one inconvenience for
another. I agree that for naive newbies, the fewer partitions, the
better,
until they gain the skill to manage multiple partitions successfully.
Newer technologies, which are rapidly gaining acceptance, may change
some of these tradeoffs, but we're not quite there yet.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/