What’s the difference between the various file system options under which to install SUSE12.2 on? Should I just stick with EXT4?
This article presents it very well. And may be of help in choosing, I’m sure you’ll get many opinions.
HTG Explains: Which Linux File System Should You Choose? - How-To Geek
On 01/10/2013 02:26 PM, Devil20 wrote:
> Should I just stick with EXT4?
that is what i use, and what i think most home users are best with…i
know btrfs sounds great, and it will be someday, but that day has not
come yet, for me…
Most filesystem choices are decent enough. The short answer I would recommend as above is to just use ext4. It’s the default for a reason. 
- It is simple and uncomplicated
- Has a long legacy of reliable technology
- Backwards compatibility
- Journaling
- Fast file system check (this really helps)
However the other filesystems can be awesome as well. The cutting edge btrfs support compression, subvolumes, and an internal raid. Though some say it has the potential for bugs being so new. I can’t say one way or the other but I personally would use something more well tested.
Xfs has traditionally awesome scalability for heavy workloads with a lot of devices. I have found it to be reliable.
Ext3 is probably about as reliable as it would get.
I am unsure if reiserfs is a real option still or if there is any advantage to it.
OK thanks chaps. I’ll stick with ext4 then.
On 2013-01-10 17:16, nightwishfan wrote:
> I am unsure if reiserfs is a real option still or if there is any
> advantage to it.
Yes to both.
–
Cheers/Saludos
Carlos E. R.
Actually, I should just assume so considering it is still in the kernel.
gnome-system-monitor says that i use ext4
Yes i did botch up and did not create a separate home partition
I wonder whether my next upgrade spare my /home 
http://i.imgur.com/w8EtQs.png](http://imgur.com/w8EtQ)
On 01/10/2013 10:16 AM, nightwishfan wrote:
>
> Most filesystem choices are decent enough. The short answer I would
> recommend as above is to just use ext4. It’s the default for a reason.
uh… but ext4 was a briefly lived apparition on SLES… and removed on SLES 11
SP2. Just saying. Ext4 is chosen in openSUSE by force, not because it’s deemed
as the “future”. We all know that it’s a temporary filesystem because we’re
sort of in limbo on alternatives. SLES 11 doesn’t seem to believe ext4 is
stable. Just saying.
With that said, I have used ext4 on openSUSE and have had no problems yet… but
do read the forums, etc… not everybody has had good experiences with ext4.
I also use XFS … but it does have some short comings… so depending on your
needs, it may be a good choice.
Red Hat (the force behind ext4 and now, sadly, xfs as well) is supporting both
ext4 and xfs as the alternatives aren’t “owned” by them…
I’m betting on the alternatives…
…snip…
>
> Xfs has traditionally awesome scalability for heavy workloads with a
> lot of devices. I have found it to be reliable.
>
> Ext3 is probably about as reliable as it would get.
Yep… like he said…
>
> I am unsure if reiserfs is a real option still or if there is any
> advantage to it.
The problem for reiserfs is support. So… while it’s superior in may ways to
current choices, it can’t be a choice just due to lack of support.
Reiserfs is still a fairly fast and easily grown (ultra fast at that)
filesystem. Sad to see it go. Try extending ext3 by 500GB and see how fast the
resize is… ext4 is faster, but still not as fast as reiserfs. That’s not
all that interesting to a home user, but in the enterprise, it’s very significant.
There’s still a glimmer of hope for reiser4, but I wouldn’t count on it.
Btrfs, now that the author is out of Oracle, has a real chance. It’s different
though. I also like the “log” filesystems like nilfs2, and others.
There’s a few ultra new players out there, but too early on…
We all know that it’s a temporary filesystem because we’re
sort of in limbo on alternatives. SLES 11 doesn’t seem to believe ext4 is
stable. Just saying.
It has been temporary for quite a long time then. 
uh… but ext4 was a briefly lived apparition on SLES… and removed on SLES 11
SP2. Just saying. Ext4 is chosen in openSUSE by force, not because it’s deemed
as the “future”. We all know that it’s a temporary filesystem because we’re
sort of in limbo on alternatives. SLES 11 doesn’t seem to believe ext4 is
stable. Just saying.
I went from ext2 to ext3 with a new install, however, on my machine ext3 had an issue of constantly accessing the disk (it never stopped), so I reverted to ext2 until ext4 was ready.
I never did find out why ext3 behaved that way on this machine, although I did search for answers, everything I found was not the cause of the issue in my case.
I have found ext4 to be totally reliable so far.
On 2013-01-11 04:46, vazhavandan wrote:
>
> gnome-system-monitor says that i use ext4
> Yes i did botch up and did not create a separate home partition
> I wonder whether my next upgrade spare my /home
Yes if you use an online upgrade or an offline upgrade - see wiki, I
don’t have the links available right now.
–
Cheers/Saludos
Carlos E. R. (12.1 test at Minas-Anor)
I wonder whether the `Swap’ file system confers a performance advantage for separate swap partitions in comparision to swap files using the ext4 file system :.
On 2013-01-11 11:26, flymail wrote:
>
> I wonder whether the `Swap’ file system confers a performance advantage
> for separate swap partitions in comparision to swap files using the ext4
> file system :.
Yes, it does. If you are using hibernation, swap partition is inmediate,
whereas a file requires the kernel to be able to read the filesystem for
read before reading swap, and this needs some heavy tinkering. I read
someone doing it, but it is something I don’t consider doing, thus I
haven’t learnt the procedure.
Otherwise a swap file seems to be as fast as a partition. I think that
the kernel maps the sectors holding the swapfile and then bypasses
filesystem routines.
–
Cheers/Saludos
Carlos E. R. (12.1 test at Minas-Anor)
On Fri, 11 Jan 2013 10:46:00 GMT, “Carlos E. R.”
<robin_listas@no-mx.forums.opensuse.org> wrote:
>On 2013-01-11 11:26, flymail wrote:
>>
>> I wonder whether the `Swap’ file system confers a performance advantage
>> for separate swap partitions in comparision to swap files using the ext4
>> file system :.
>
>Yes, it does. If you are using hibernation, swap partition is inmediate,
>whereas a file requires the kernel to be able to read the filesystem for
>read before reading swap, and this needs some heavy tinkering. I read
>someone doing it, but it is something I don’t consider doing, thus I
>haven’t learnt the procedure.
>
>Otherwise a swap file seems to be as fast as a partition. I think that
>the kernel maps the sectors holding the swapfile and then bypasses
>filesystem routines.
It thoroughly does bypass other filesystems. It is in many respects a
version of virtual memory. It has other uses for sleep, suspend, and
hibernate. If you want to use these methods (sl, sus, & hib) having swap
larger than ram is a good idea.
?-)
Approx Oct 2012 I did some personal research in prepararion for a new install on a laptop upgraded with an SSD.
At the moment only ext4 supports SSD functions like TRIM natively. With any other filesystem you’d have to install a utility to perform those essential functions and configure a cron job to run automatically.
Although there are/were suggestions that ext4 is only a stepping stone to something far more advanced like btfrs, ext4 has nicely proven relliable, extensible and updateable with features like TRIM. Also, although next gen fs might already have various features built-in, it’s comforting that so many familiar fs utilities that worked on ext2 and ext3 can in most cases be easily updated to run on ext4.
So, in general the most reliable with useful features is today ext4. For those who can use special features of xfs, is a good alternative. There are few reasons to consider another fs including earlier ext, eg specialized hardware(I’ve noticed QEMU openSUSE ARM selects ext3 by default)
HTH,
TSU