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

Re: XFS, EXT3 or some other?



Dan Serban <dserban@lodgingcompany.com> writes:

> Hemlock wrote:
>> Hello,
>> Seeking some advice on which filesystem to use.
>> I've been primarlity an ext3 user, because thats all I've really
>> ever known of, but would consider trying something else.
>> Plus, never used anything other than 32bit Debian since slink.
>> 
>> I've read some articles googling for xfs, ext3 and jfs and such.
>> Leaning towards xfs maybe?
>> 
>> This will be for a AMD64 sarge box used as a file, web, email and proxy 
>> server. Also running software raid.
>> 
>> Thanks for your advice.
>> 
>> 
>
> I can't much comment on the different options available as I've moved
> from ext3 to jfs recently (say, for the last 6 months).
>
> My primary reason from switching away from ext3 were the long fsck's on
> boot after a disatrous crash or power outage.  I then researched what I
> could between ext3, xfs, jfs, and reiser.  Here's what I found (this is
> all from memory and personal opinion).

Ext3 does only a journal recovery after a crash and that only takes a
few seconds. Ext2/3 does a full fsck every X mounts or Y days (both
configurable) which you might hit every time if you reboot as seldom
as me.

> ext3: not optimized in many ways compared to the other journaling file
> systems available, long periods of recovery and mild corruption, though
> on a whole, it has a good reputation on data integrity.  And the issues
> I've mentioned seem to be worked on in more recent kernel revisions.
>
> XFS: painful in some ways (can't remember specifics, but in one case you
> can thoroughly hose your filesystem) though, xfs was one of the choices
> I seriously looked at, I don't remember the specifics, but such small
> things such as long time to remove files, and other minor issues such as
> CPU utilization.  Though I'd try this one next if jfs lets me down.

Hmm, ext3 is slow to delete since it deletes every data block from the
inode structure and all inodes from the indirect inodes in turn and so
on. For larger file that means quite a few commits through the
journal. It also makes undelete of files impossible.

Never heard that about xfs.

MfG
        Goswin



Reply to: