Re: RAID and preferable file systems
Mark L. Kahnt said:
> No - for me it means Redundant Array of Inexpensive Drives - with
> emphasis on the redundancy and the ability to replace (hopefully
> hotswap) and rebuild/maintain data availability. I've done a couple of
> servers with RAID1 running ext3, but I haven't actually seen anything
> about good/bad filesystems in that environment, and I am looking at a
> project which leans to RAID5, which has me trying to find useful
> information before bidding on the installation (aka, is it going to be
> more of a pain than it is worth?)
it shouldn't matter. hardware RAID should be transparent to the OS.
I have run hardware raid 5(SCSI) on ext2 and have 4+1 RAID5 array
at home with redhat 7.3 on ext3. Also have run hardware RAID 5(IDE)
on reiserfs, hardware RAID 10(IDE) on reiserfs. Hardware raid 1(IDE)
on reiserfs & ext2. Software raid 1(2.2.x kernel) on ext2, software
raid 1(2.4.x kernel) on reiserfs, also software raid 0 on ext2.
the level of raid really doesn't matter. if anything the application
that the array will be storing will matter, will it deal with many
tiny files or a small number of large files? reiserfs seems to be
better(from what I read) for systems that have lots of tiny files though
in my experience I cannot tell a difference between ext3 and reiserfs,
I lean towards reiserfs when I can since I've worked with it more then
I have with ext3 and it seems to work well for me.
ext3 may be better in some recovery situations as you can sometimes
access the filesystem in ext2 "mode" (provided the journal is intact?)
and it provides an easy upgrade path from ext2(and vise versa I think?)
so, again. filesystem shouldn't matter when your focusing on what
RAID level that you'll be using.
nate
Reply to: