[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Ext3 or XFS (or none)?



<quote who="Vineet Kumar">
> * nate (debian-user@aphroland.org) [020709 14:22]:
>> it is not an easy migration path, unlike ext2->ext3, but for me
>> i'd rather just start fresh then risk leaving old cruft in
>> the filesystem that was a result of a conversion.
>
> Well, if by "old cruft in the filesystem" you mean "the filesystem", then
> you're right ;-) The ext2->ext3 conversion isn't even really a
> conversion. It's just adding a journal to an existing ext2 filesystem.
> *Everything* that was in the filesystem originally is left there.
>
> Incidentally, you should get no more or less than if you did a full
> backup, create a reiserfs, and restored the backup. (Note, I said "full
> backup!") So I'm not sure what you mean by this at all. It sounds like
> you're advocating a system reinstall instead of a backup/restore, but
> that shouldn't be necessary unless the admin has done something bad.
> (i.e. an accidental chmod -R on / or the like). Even still, you have
> backups, right? =)

well i just meant that sometimes when converting between systems
the conversion doesn't do as clean of a job as just starting fresh.
maybe ext3 does..ive read that ext3 is not much more then ext2
just a journal file and some things, and the filesystem is
still accessable to the ext2 driver ..


and yeah i would do a full backup/restore on a fresh filesystem
unless i didn't have the means to do so. im just paranoid i guess.
*shrug*

if i didn't have the means, i'd not do the the change, or wait
till i had the means ..

if the system has a good partition scheme (different partitions
for /usr /var / /boot /home), it should be pretty easy to migrate.
infact i did this from remote once, was pretty scary, i changed
runlevel 3's init scripts to drop to runlevel 1, format the
new partitions, make new mount points, mount the drives, copy
the data, rename the old directories (usr->usr.old etc),
unmount the drives, make new mountpoinuts(/usr /var etc),
remount the drives on the new permanent mountpoints, put
entries in /etc/fstab, and come back to runlevel 2.

took about an hour of testing, and when i hit the button it
was a real long minute or 2!! but it came back! and it
was nearly perfect(i think i screwed up the fstab..but
thats easy to fix). thanks to vmware it was easy to test.

all my systems have real small / partitions with the
exception of those that boot off of software raid 1 with
/ on the raid array ..usually 500Mb or less

nate




-- 
To UNSUBSCRIBE, email to debian-user-request@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org



Reply to: