What is Digg?Sponsored by Newegg.com
$5,000 Newegg Eggstreme Challenge Video Contest Extended to 11/15 view!
newegg.com - Show them what YOU would do with the money you save by shopping at Newegg. Seriously, show them...
48 Comments
- trafficlight, on 04/29/2009, -1/+21It's been murdered.
- overnine9k, on 04/29/2009, -1/+17So no more ReiserFS?
- SteveMax, on 04/29/2009, -1/+15Everything around Reiser seems to be dying nowadays.
- tofagerl, on 04/29/2009, -0/+12How's BtrFS doing? Genuinely curious if anyone has heard any news since Oracle bought Sun and got ZFS. A GPLed BtrFS is better than a non-GPLed ZFS :(
- Megatog615, on 04/29/2009, -0/+11Because FAT, FAT32, and NTFS don't respect UNIX file permission schemes.
- nirvanix, on 04/29/2009, -0/+8ReiserFS 3 is an excellent file system, and version 4 has been reported to be excellent. Are they dropping it because Hans Reiser is in prison?
- Celarnor, on 04/29/2009, -1/+8Really? I've always experienced the exact inverse of this right out of the box. None of my Thinkpads, my HP, or my old Compaq got better battery life in Windows.
- GameDNA, on 04/29/2009, -0/+6without someone to lead the project, it seems to be dying...
- Phocion55, on 04/29/2009, -0/+4I'll rephrase it.
Every other section on this website other than the Powertop section may help you: http://www.lesswatts.org/
(site isn't specific to just powertop)
And no, I don't read every comment word-for-word. - evilxhwnd, on 04/29/2009, -3/+8you need to do more testing then. seriously.
- BadLittleGuy, on 04/29/2009, -0/+3If you don't know them:
posix acls. Namely look into setfacl and getfacl.
If you do know them:
What's really missing? - lyonsban, on 04/29/2009, -0/+41) don't blame me if you hose your computer
2) RTM for fstab options, umask, uid & gid - failtrain, on 04/29/2009, -1/+5It is pretty much dead and buried now.
- techdever, on 04/29/2009, -2/+6"even after doing powertop tweaks" wow.. can't you people read before you digg a comment down ?
- adriaaan, on 04/29/2009, -0/+3Ext2/Ext3
Windows: http://sf.net/projects/ext2fsd/
Mac OS X: http://sf.net/projects/ext2fsx/ - Phocion55, on 04/29/2009, -1/+4This may help you: http://www.lesswatts.org/projects/powertop/
- redled, on 04/29/2009, -0/+3I couldn't find any new news. I'm using ZFS on FreeBSD and I hope its stable and usable in linux via fuse soon. It's very nice.
- SteveMax, on 04/29/2009, -0/+2That's a mount option. Your distro may be passing "users" as the default to mount (which gives access to all users), while you want to pass "user" (which only gives access to the mounting user).
Again, it is possible that the specific FS driver you are using (as in "the specific NTFS implementation you use") doesn't understand "user", but if that is the case, the driver is outside the kernel. - BadLittleGuy, on 04/29/2009, -0/+2Look again at the documentation of ntfs-3g. It actually supports mapping of unix account to windows ones. It's just a huge pain in the ass, because unix and windows permission systems are completely different, not to mention finding the right windows uids to map to.
- lyonsban, on 04/29/2009, -0/+2"umask" sets your file/directory permissions for the mount.
"uid" and "gid" control ownership and group of the mounted files. - Darkhacker, on 05/01/2009, -0/+2>"it still one of the desktop usable filesystems vastly outperforms ext3 (the other being jfs) and offers more stability than ext4/btrfs."
Incorrect. Ext3 is either on-par or exceeds Resier3 in performance tests.
source: http://www.phoronix.com/scan.php?page=article& ...
Reiser4 is another story. It performs very well, but is not in the mainline kernel (only Andrew's -mm branch). The reason it's not in mainline is that certain coding standards aren't followed by Reiser4. Plus it's still under development and I wouldn't quite trust it in a stable environment.
JFS does outperform ext3 but with a couple of gotchas. The first being that it requires an fsync on rename. In other words, it acts like ext4 did before the Tso's patches. I can attest to this as I've lost data with JFS. The second being that you cannot shrink a JFS partition. Granted you wouldn't do this often, but it's still annoying. I've also heard that JFS becomes slow with directories that have a lot of files in them, but I haven't found a benchmark that confirms this, only user testimonies. All in all though, I'd rather have the safety ext3 provides. Especially since I have lost data with JFS and have never lost a thing with ext3. - MikeVx, on 04/30/2009, -0/+2It's still there, just lacking in development now. Without a maintainer it just gets tweaks and whatever adjustments any code gets when the kernel changes, and the odd fix that someone submits. While it will probably go away eventually, I expect that to be a few years down the line. I haven't let the whole murder issue drive me to a change, I still use Reiserfs. At the time I started using it, it was the only readily available file system that could do files larger then 2GB. The others were hairier to set up. Unless another fs turns up that addresses some problem I have, I'll probably use Reiserfs until it gets deprecated, then I'll figure out what to use next.
- inactive, on 04/29/2009, -0/+2http://www.flickr.com/photos/ronocdh/2651155110/si ...
- inactive, on 05/01/2009, -0/+2Reiser3 is not dead OR buried.
Reiser3 wasn't under development for quite some time BEFORE hans went and got himself locked up, it still one of the desktop usable filesystems vastly outperforms ext3 (the other being jfs) and offers more stability than ext4/btrfs. The novel guys have been patching stuff here and there but going into 'deep' maintenance isn't really a significant change. - cesclaveria, on 04/29/2009, -1/+3different options and different needs, not just because something already works means all work on that field must be stopped.
Firefox 2 was a great browser, but that did not means that firefox 3, chrome, ie 8, etc are welcome additions to the browser market. Now apply the same principles here and for filesystems is probably even more critical, since the hardware is advancing really fast, the way the people uses computers change, the kind (and size) of files that the users have now are different and will keep changing. - tomz17, on 04/29/2009, -0/+1I've been using XFS for the past few years, and love it. Performance is much better than ext3, handles large files and filesystems very well, and even has an online defrag utility! In fact, I see very little reason to switch to ext4 from xfs.
- philz, on 04/29/2009, -0/+1I guess you forgot to clear your cache.
- inactive, on 04/29/2009, -0/+1As a former ReiserFS user who 'headed for the hills' when Reiser was arrested, the kernel team would save themselves some work if they abandoned it. Nobody in their right mind would use it at this point. Sad to hear developers are still making excuses for ext4. It screams, "Stay away!"
- paulsmith288, on 04/30/2009, -0/+1My battery life has been about the same with windows and linux.
- Lamnk, on 04/29/2009, -0/+1I don't know why you are down voted.
Have the same experience: running Archlinux, kernel 2.6.29 gives me 2 hours battery max. With windows XP it is 2h 30min. Go figure. - tofagerl, on 04/29/2009, -0/+1I'm also using ZFS on FreeBSD/FreeNAS. A little sketchy tbh, a lot of kernel panics related to ZFS. I'm considering switching to FreeBSD "pure", but I fear the configuration nightmare, which is so easy with FreeNAS.
- techdever, on 04/29/2009, -1/+2I didn't say anything about thinkpads, they are great in linux. What HP do you have ?
- lyonsban, on 04/29/2009, -1/+2NTFS is such a horrible file system that the penalty in the kernal for supporting a translation table from POSIX to Windows ACLs wouldn't be worth it. You'll never get the kernal team to agree to the penalty. At the application level you can work around it "fairly" easily, such as Samba does. I would like to see better ACLs above and beyond those in POSIX though, maybe it'll be in 2.8, if IEEE agrees.
- techdever, on 04/29/2009, -3/+3yeah, cause i have all the laptops in the world to test
/s - lyonsban, on 05/01/2009, -0/+0I forgot about that. You're right. I ended up using linux permissions, for that reason. The uids are easy enough to find and use, but keeping it updated never mind synced with windows almost needs a new employee to keep it current. I just ditched windows for everything except business applications and used samba for everything else from file servers, domain controller and printing.
- wicketr, on 04/29/2009, -1/+1Well, if NTFS stinks, what file system should I use so that it can be read by both Windows and Linux, AND have file permissions obeyed by both OSes?
As far as I can see, NTFS is the most acceptable answer to this question. - wicketr, on 04/29/2009, -2/+2Well, ok. So do we just give up? They need to create a security layer that lies between them. Surely they can understand the permissions that Windows has implemented for at least some basic security.
And as for me mounting the drive (password protected mount), I don't understand why Linux passes the access to other users. It should be a user by user basis unless specifically stated. If I type in the password to mount the drive, then i should be the only one that can access it. Not my roommate who comes in 2 hours later.
For an OS and developers that are so anal retentive about users needing to type their name in at the login as well as the password (for fear of someone guessing the username), this seems like a huge security hole. - SteveMax, on 04/29/2009, -3/+2I see them all the time.
http://en.wikipedia.org/wiki/Faddeev-Popov_ghost - srg13, on 04/30/2009, -2/+1Does it really make you that upset that you can't understand the article that you have to make stupid comments about it?
- SchrodingersCar, on 04/29/2009, -3/+1UGH! I'm an idiot. I obviously meant P-Dex 3 not P-Dex 4 (since it does not yet exist...) You are right to digg me down. Way to proofread, numbskull!
- MattS, on 04/29/2009, -4/+2wow, you sure do sound smart.
- Opperior, on 04/29/2009, -1/+0I would like to see better support for ACLs, myself. POSIX-style file permissions are a bit too restrictive for my tastes.
- techdever, on 04/29/2009, -8/+5Still waiting for a kernel that gives battery life at least comparable to vista on laptops. Linux (most distros) even after doing powertop tweaks still falls way behind at battery life on most laptops I've tested on (apple,sony, dell and hp)
- wicketr, on 04/29/2009, -5/+2Why don't they fix NTFS support so that it actually obeys NTFS permissions?? I hate the fact that I mount the drive on my account and then every other user also has access unless i remember to unmount it every time. Annoying.
- davzie, on 04/29/2009, -5/+1Yeah seriously, I think I'm going to dig you down simply because it sounds far too smart to be true
*shrugs* - Rhino2, on 04/29/2009, -10/+2Serious question, but why is Linux always dicking around with file systems? That's all I see "Hey, guys we got a new file system... it's better than the last". Don't have have anything else to work on?
also
> Although it is now barely maintained, there are also fixes for ReiserFS
Seriously, no one wants a half-assed and half-supported FS in their god dam OS. THIS is an example of the reasons why people don't take linux seriously. - SchrodingersCar, on 04/29/2009, -11/+2It's about time we see some more robust implementation of PDoT and TDoT revisions as well as Tri-X support for the kernel - not only for security and Ext4, but to make sure developers who have spent way too much time (and wasted P-Dex 4 resources) trying to hash-on-the-fly working in RTOFS or similar, no longer get the short end of the stick when it comes down to the brass tacks of RegCOR 2.0 and Openbyte integration in the 4-1 to 8-1 range.
© Digg Inc. 2009
— Content created and posted by Digg users is