LinuxQuestions.org
Download your favorite Linux distribution at LQ ISO.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices

Reply
 
Search this Thread
Old 03-15-2013, 05:25 PM   #16
kikinovak
Senior Member
 
Registered: Jun 2011
Location: Montpezat (South France)
Distribution: Slackware, Slackware64
Posts: 1,489

Rep: Reputation: 662Reputation: 662Reputation: 662Reputation: 662Reputation: 662Reputation: 662

Mainly ext4 on pretty much everything, with the exception of ext2 for my small /boot partition, where there's no need for a journaled filesystem.
 
Old 03-16-2013, 12:05 AM   #17
rkelsen
Senior Member
 
Registered: Sep 2004
Distribution: slackware
Posts: 1,748

Rep: Reputation: 159Reputation: 159
xfs. Because it kicks butt.

I've been using it since 2006 and never had a problem. I can't tell if it runs an fsck at boot, because fsck on xfs takes milliseconds.

It doesn't pollute partitions with ugly 'lost+found' directories.

Before then I used reiserfs and had done so for 6 or 7 years without any problems. When Hans Reiser was charged with murder, I wanted to try something different. I tried jfs, but found it slower than ext filesystems. xfs is fast, responsive and has excellent recovery capabilties.
Quote:
Originally Posted by tronayne View Post
Here's something to mull over: Slackware defaults to ext4, Pat and crew do that, why?
Uhhh... No, slackware doesn't default to any filesystem. In fact, the installer asks you which one you want to use. You still need to build an initrd to use a generic kernel with an ext4 root.

Last edited by rkelsen; 03-16-2013 at 12:07 AM.
 
1 members found this post helpful.
Old 03-16-2013, 12:33 AM   #18
JWJones
Member
 
Registered: Jun 2009
Location: Cascadia
Distribution: Slackware, LinuxBBQ, OpenBSD, Mac OSX
Posts: 722

Original Poster
Rep: Reputation: 184Reputation: 184
Quote:
Originally Posted by rkelsen View Post
Uhhh... No, slackware doesn't default to any filesystem. In fact, the installer asks you which one you want to use.
Yeah, now that you mention it, this is true. And this would be in keeping with the Slackware way, of course.
 
Old 03-16-2013, 12:49 AM   #19
ttk
Member
 
Registered: May 2012
Location: Sebastopol, CA
Distribution: Slackware
Posts: 175
Blog Entries: 10

Rep: Reputation: 85
Working at archive.org from 2004 to 2008, sometimes I had to deal with a great many files (often thousands, sometimes millions, occasionally tens of millions). I quickly became enamored of XFS for its robustness when dealing with single large files and oodles of small ones.

XFS also comes with a multitude of handy administrative tools: xfs_repair xfsdump xfsrestore xfs_admin xfs_bmap xfs_check xfs_copy xfs_db xfs_estimate xfs_freeze xfs_fsr xfs_growfs xfs_info xfs_io xfs_logprint xfs_mdrestore xfs_metadump xfs_mkfile xfs_ncheck xfs_quota xfs_rtcp xfsdump xfsinvutil xfsrestore xfsinfo

I almost never need them, but when I do, they are very nice to have.

A lot of my techie friends are enamored of ext4, and have had very nice things to say about it, but I was very happy with xfs and saw no reason to change.

Recently, however, I've started to worry about the xfs development team. They're not sponsored anymore, and not really thriving. They recently had to perform major surgery on the code to close a performance gap with ext4, and I worry that they haven't had the engineering-hours to spare to properly QA the code.

I get the impression that the xfs project is in slow decline, which is unfortunate, because it's been working really well for me.

I'm still using xfs as my workhorse filesystem on most of my computers, but when I got a new hard drive for my home desktop, I went ahead and formatted it ext4. I just want to have one in my life so I can learn my way around it, perhaps see a failure mode and learn how to deal with it. That way if xfs ever ceases to be a valid option, I won't be completely in the dark when transitioning to ext4.

That's perhaps some ways into the future, however. For now, xfs continues to work very well. It lacks the scale problems of ext[2,3] and reiserfs. It has no /lost+found because it never loses track of blocks. Unmounting it uncleanly does not require a fsck before it can be mounted again, because it performs the equivalent of fsck constantly during its normal operation. If it can keep its head above water, I wouldn't mind using it indefinitely.
 
4 members found this post helpful.
Old 03-16-2013, 01:19 AM   #20
p2006.prashant
LQ Newbie
 
Registered: Dec 2007
Location: India
Distribution: Redhat Linux 5,6.2, Solaris
Posts: 16
Blog Entries: 1

Rep: Reputation: 5
ext4 because it is latest version which has many more features than earlier versions.
 
Old 03-16-2013, 04:39 AM   #21
chrisretusn
Member
 
Registered: Dec 2005
Location: Philippines
Distribution: Slackware
Posts: 465

Rep: Reputation: Disabled
Quote:
Originally Posted by rkelsen View Post
Uhhh... No, slackware doesn't default to any filesystem. In fact, the installer asks you which one you want to use. You still need to build an initrd to use a generic kernel with an ext4 root.
I guess it depends on your definition of "default". The default choice is ext4, if you hit enter then ext4 is what you get, by default. If you choose to use the default huge kernel, then nothing more needs to be done.

To stay on topic. I use prefer to use jfs, No real reason aside from that fact I have been using it for years. I also use ext4 on some of my installs The jfs file system does well with huge files, a draw back of jfs is file deletes, which can take a bit of time for a large number of deletions.

I did run in to a problem with ext4 and corrupted files a few years back. It was part of a troubleshooting effort with others. I was not seeing any corruption problems, I was using jfs. Switching over to ext4 resulted in corruption. It is my understanding this problem with ext4 has been since been fixed. Just the same, if I am dealing with large files and databases I prefer not to use ext4.

Last edited by chrisretusn; 03-16-2013 at 06:02 AM.
 
Old 03-16-2013, 05:37 AM   #22
kooru
Senior Member
 
Registered: Sep 2012
Location: Italy
Distribution: Slackware, NetBSD
Posts: 1,230
Blog Entries: 4

Rep: Reputation: 256Reputation: 256Reputation: 256
ext4. no problem with it.
 
Old 03-16-2013, 06:34 AM   #23
rkelsen
Senior Member
 
Registered: Sep 2004
Distribution: slackware
Posts: 1,748

Rep: Reputation: 159Reputation: 159
Quote:
Originally Posted by chrisretusn View Post
I guess it depends on your definition of "default". The default choice is ext4, if you hit enter then ext4 is what you get, by default.
Referring to that as any kind of endorsement seems to be a bit of a stretch... but OK...
Quote:
Originally Posted by chrisretusn View Post
If you choose to use the default huge kernel, then nothing more needs to be done.
The same can be said for any other filesystem you choose, since the huge kernel has support all of the abovenamed filesystems compiled in - and more.
Quote:
Originally Posted by chrisretusn View Post
I did run it do a problem with ext4 and corrupted files a few years back.
Funny you mention that. I had something similar with ext2, back in the day. The endless fscks on every 10th boot of my desktop machine drove me nuts. Back then you had to patch the kernel if you wanted something like reiserfs. I remember running Slackware 7.0 with a highly custom 2.2 kernel.
 
Old 03-16-2013, 09:25 AM   #24
chemfire
Member
 
Registered: Sep 2012
Posts: 68

Rep: Reputation: Disabled
BTRFS all the way snapshots are to useful to live without anymore.
 
Old 03-16-2013, 09:31 AM   #25
irgunII
Member
 
Registered: Jan 2012
Location: Directly above the center of the earth
Distribution: Slackware. There's something else?
Posts: 363

Rep: Reputation: 61
I've been using ReiserFS since I began using Linux back in 2000. It's my opinion that my hdd's and thus all the info on them has been saved/safe from the literally countless times we get brown/blackouts here. It's fast enough for my needs (I'm just a plain user, nothing fancy, can work on .conf files but can't write a batch script) and like another poster stated earlier, it simply works, you don't fix what ain't broke.

As for those who say it's unstable...how so? Qualify your statements rather than just spew them out there and expect the world to think it's the only correct opinion there is. If it was so unstable, why is it still offered with almost every distro?
 
Old 03-16-2013, 10:03 AM   #26
H_TeXMeX_H
Guru
 
Registered: Oct 2005
Location: $RANDOM
Distribution: slackware64
Posts: 12,928
Blog Entries: 2

Rep: Reputation: 1269Reputation: 1269Reputation: 1269Reputation: 1269Reputation: 1269Reputation: 1269Reputation: 1269Reputation: 1269Reputation: 1269
See here for criticism of ReiserFS:
http://en.wikipedia.org/wiki/Reiserfs#Criticism
I think that ReiserFS is outdated and development has stalled for obvious reasons.

I forgot to mention that JFS works best with the deadline I/O scheduler from my experience and from the arch wiki:
https://wiki.archlinux.org/index.php....2FO_Scheduler
 
Old 03-16-2013, 10:49 AM   #27
allend
Senior Member
 
Registered: Oct 2003
Location: Melbourne
Distribution: Slackware-current
Posts: 3,378

Rep: Reputation: 824Reputation: 824Reputation: 824Reputation: 824Reputation: 824Reputation: 824Reputation: 824
I have a legacy install that uses reiserfs, but all later installs are ext4.
I also disagree that reiserfs is unstable. It has always successfully recovered from power disruptions. My impression, although I have no hard data, is that it is better for accessing large numbers of relatively small files (250K to 8MB) in succession than ext4.

PS- Happy St Patrick's day to all!

Last edited by allend; 03-16-2013 at 10:50 AM.
 
Old 03-16-2013, 01:41 PM   #28
ttk
Member
 
Registered: May 2012
Location: Sebastopol, CA
Distribution: Slackware
Posts: 175
Blog Entries: 10

Rep: Reputation: 85
Quote:
Originally Posted by irgunII View Post
As for those who say it's unstable...how so? Qualify your statements rather than just spew them out there and expect the world to think it's the only correct opinion there is. If it was so unstable, why is it still offered with almost every distro?
When I joined Discovery Mining in 2008, they were in the process of transitioning away from ReiserFS on their cluster of CentOS servers. Their main problem with it was, when deleting many files (either hundreds or thousands, I don't remember), the system became completely unusable for a long time while ReiserFS performed the deletions.

Their workaround was to fix up our data processing software so that, instead of deleting files, files were moved to a special directory ("/var/TO_BE_DELETED", I think). Moving the files was a very fast operation. Then a cron job would delete the contents of this directory in the wee hours of the morning, when it was assumed there would be less demand on the cluster.

This bit them in the ass a few times when customers, who needed to work through the night in order to make a deadline, ran into slowness using our service when the deletions were taking place. It was decided that ReiserFS had to go.

ext4 was not considered a viable candidate for production use in 2008, so they ran a benchmark approximating our peak workload to decide between ext3 and xfs. I wish I'd gotten there sooner, to influence the process more, because I knew from past experience that ext3 did *not* scale well to many files. Unfortunately they chose ext3, because it beat out xfs by a slim margin in their benchmarks, and I had to deal with that decision for the following three years.

Eventually Discovery Mining was purchased by Autonomy, which was really horrible for the company. Employees left in droves. Because of where I live, it was harder for me to find a place to jump ship, so I ended up staying long enough to be one of the only two engineers capable of running, developing, and maintaining the system. This gave me more leeway to rearrange, and I reformatted several filesystems to xfs. Those servers I used when I had to manipulate dataloads in such a way that ext3's 32K link limit posed a liability. As usual, xfs took it like a champ.
 
Old 03-16-2013, 06:26 PM   #29
D1ver
Member
 
Registered: Jan 2010
Distribution: Slackware 13.37
Posts: 527
Blog Entries: 3

Rep: Reputation: 126Reputation: 126
Quote:
Originally Posted by ttk View Post
Recently, however, I've started to worry about the xfs development team. They're not sponsored anymore, and not really thriving. They recently had to perform major surgery on the code to close a performance gap with ext4, and I worry that they haven't had the engineering-hours to spare to properly QA the code.

I get the impression that the xfs project is in slow decline, which is unfortunate, because it's been working really well for me.

I'm still using xfs as my workhorse filesystem on most of my computers, but when I got a new hard drive for my home desktop, I went ahead and formatted it ext4. I just want to have one in my life so I can learn my way around it, perhaps see a failure mode and learn how to deal with it. That way if xfs ever ceases to be a valid option, I won't be completely in the dark when transitioning to ext4.

That's perhaps some ways into the future, however. For now, xfs continues to work very well. It lacks the scale problems of ext[2,3] and reiserfs. It has no /lost+found because it never loses track of blocks. Unmounting it uncleanly does not require a fsck before it can be mounted again, because it performs the equivalent of fsck constantly during its normal operation. If it can keep its head above water, I wouldn't mind using it indefinitely.
I switched the filesystem on my new laptop and media server over to XFS after watching this linuxconf talk on XFS. He mentions that XFS has caught up to ext4 in areas where ext4 was traditionally stronger, while keeping it's own strengths. That combined with a quote from ext4 developer Theodore Ts'o where he mentioned that ext4 is a stopgap on the way to BTRFS had me move away from ext4..
 
1 members found this post helpful.
Old 03-16-2013, 07:11 PM   #30
rg3
Member
 
Registered: Jul 2007
Distribution: Slackware Linux
Posts: 504

Rep: Reputation: Disabled
ext4 on an SSD, because it supports TRIM and is used by lots of people, so it's very well tested and supported.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
How to use key file instead of password for LUKS encrypted file systems? lucmove Linux - Security 2 06-30-2009 09:17 AM
LXer: Kernel Log: What's coming in 2.6.30 - File systems: New and revamped file syste LXer Syndicated Linux News 0 04-27-2009 11:02 PM
Read only file systems, custom live CDs, and embedded systems coffeecoffee Linux - Newbie 2 02-24-2009 11:09 PM
LXer: Cluster File Systems Attains World Leadership Position In High Performance File System LXer Syndicated Linux News 0 07-14-2006 06:21 AM
Is arrangement of file systems will differ if we copy a file from FAT 32 to ext 3 ? anindyanuri Linux - Software 2 02-20-2005 11:39 AM


All times are GMT -5. The time now is 09:20 AM.

Main Menu
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration