LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Slackware (https://www.linuxquestions.org/questions/slackware-14/)
-   -   ReiserFS problems solved by switching to ext4 (https://www.linuxquestions.org/questions/slackware-14/reiserfs-problems-solved-by-switching-to-ext4-4175425294/)

EdGr 09-03-2012 02:21 PM

Quote:

Originally Posted by durval (Post 4771674)
Ditto ;-) I expect to be in your situation when ZFSOnLinux stabilizes enough (and has good enough performance) for me to migrate to it; with any luck I will be able to bypass ext4 completely and (I hope) migrate to BtrFS if and when it achieves a reasonable level of maturity.

I initially resisted moving to ext4 because the code has a lot of cruft. I thought that ext4 would be a step backwards from ReiserFS, which is quite elegant.

I'm now finding that the crufty working filesystem is infinitely preferable to the elegant broken one. :)
Cheers,
Ed

durval 09-03-2012 02:52 PM

Hi EdGr,

Quote:

Originally Posted by EdGr (Post 4771690)
I initially resisted moving to ext4 because the code has a lot of cruft. I thought that ext4 would be a step backwards from ReiserFS, which is quite elegant.

My problem with ext4 is the same as with ext3: in my experience, it's not a reliable file system if things go wrong. Even simple things like a lockup/panic, or an unexpected poweroff, has a high probability of corrupting the filesystem in a way that it can't recover by itself on the next boot (and to think that a *log* structured filesystem was supposed to solve almost exactly that problem... :-/). Reiserfs, on the other hand, has had its (virtual) power plug yanked from it hundreds of times in my tests (resetting the VM where I'm running it) and it has never ever lost any data, or been incapable of booting.

My problem with btrfs for now is that it's simply too immature (the patch descriptions that people are constantly submitting reads like a bizantine horror story)... I do hope that the developers manage to stabilize things quickly and also that they change its stupid (IMNSHO) behavior regarding the handling of free space vs erased files... the current behavior is very unintitive and manages to crash my tests using dledford-memtest (and other stressing programs that I use to validate an installation) due to "out-of-disk-space" conditions, even though (when you account for the files that have been removed) there's still a whole lot of space left.

Quote:

I'm now finding that the crufty working filesystem is infinitely preferable to the elegant broken one. :)
LOL! You are so right... ;-) in fact, I think I will copy your phrase to our local "fortune" file... how do you want to be cited?

Cheers,
--
Durval.

yenn 09-03-2012 03:06 PM

I wanted to use ReiserFS on new servers I'm going to install very soon, because I use ReiserFS almost exclusively from 2004. But now I see that while it works flawlessly on my single-cpu machines (don't have any multi-core around), things may go haywire on much newer hardware. So I'll give ext4 try. Thanks for this thread and suggestions.

EdGr 09-03-2012 03:07 PM

Quote:

Originally Posted by durval (Post 4771712)
My problem with ext4 is the same as with ext3: in my experience, it's not a reliable file system if things go wrong. Even simple things like a lockup/panic, or an unexpected poweroff, has a high probability of corrupting the filesystem in a way that it can't recover by itself on the next boot (and to think that a *log* structured filesystem was supposed to solve almost exactly that problem... :-/). Reiserfs, on the other hand, has had its (virtual) power plug yanked from it hundreds of times in my tests (resetting the VM where I'm running it) and it has never ever lost any data, or been incapable of booting.

Agreed: ReiserFS was very good at surviving power outages and crashes. Thanks for the warning about ext4 - I haven't been using it for long enough to find out how robust it is.

Quote:

Originally Posted by durval (Post 4771712)
LOL! You are so right... ;-) in fact, I think I will copy your phrase to our local "fortune" file... how do you want to be cited?

EdGr at linuxquestions.org. Thanks!
Ed

durval 09-03-2012 03:55 PM

Hi EdGr,
Quote:

Originally Posted by EdGr (Post 4771721)
Agreed: ReiserFS was very good at surviving power outages and crashes.

As far as I'm concerned (and while ZFSOnLinux and btrfs don't become viable options), ReiserFS still is :-)

Quote:

Originally Posted by EdGr (Post 4771721)
Thanks for the warning about ext4 - I haven't been using it for long enough to find out how robust it is.

You are welcome. Resisting power outages is really important for us here because some of our clients have really bad electrical infrastructure and additionally can't afford an UPS for every machine, specially when they are widely dispersed in a plant. And resisting crashes is important for our development, as we do crazy server-crashing stuff frequently and it sucks having to reinstall an entire machine from backup because the filesystem just went poo-poo on you... :-)

Cheers,
--
Durval.

Pixxt 09-03-2012 10:49 PM

Quote:

Originally Posted by EdGr (Post 4771721)
Agreed: ReiserFS was very good at surviving power outages and crashes. Thanks for the warning about ext4 - I haven't been using it for long enough to find out how robust it is.

EdGr at linuxquestions.org. Thanks!
Ed

LOL I do not want to laugh but ReiserFS is among the least robust filesystem I have ever used. I had nothing but problems with ReiserFS, as i said before lame ass FAT(16/32) was a better FS for stability in my experience than reiser.

ReaperX7 09-03-2012 11:02 PM

Unless ZFS, or at least ZFSOnLinux gets forked gets re-licensed to maybe MIT, or Linux gets relicensed to something other than GPL, like MIT, that's more open to other open-source licenses, we'll never get ZFS on Linux officially.

It's a shame because ZFS has withstood the test of time for over 10 years now and has been the premiere of UNIX file systems. heck BtrFS can barely mimic and has yet perform the same functions as ZFS.


All times are GMT -5. The time now is 07:20 AM.