Debian Planet










Welcome to Debian Planet

Search

All your woody are (not quite, but very very very soon) belong to us.
Main Menu

  • Home

  • Topics

  • Web Links

  • Your Account

  • Submit News

  • Stats

  • Top 10

  • Debian

    These are important Debian sites one should not be without!

  • Official Debian site

  • Package search

  • Mailing list archives

  • Bug reports

  • Debian on CD

  • Unofficial woody CD ISOs

  • Unofficial APT sources

  • Developers' Corner

    Other great Debian news sources:

  • Debian Weekly News

  • Kernel Cousin Debian

    (Debian mailing lists digested)
  • Community Groups

    Need help? You're not alone on this planet.

  • debianHELP

    (User support site)

  • Debian International

  • DebianForum.de

    (Deutsch)

  • EsDebian

    (español)

  • DebianWorld

    (français)

  • MaximumDebian

    (Italiano)

  • DebianUsers

    (Korean)

  • Debian-BR

    (Português)

  • IRC

    The place to get help on a Debian problem (after reading docs) or to just chat and chill is #debian on irc.debian.org.

    Many of the Debian Planet staff live there so pop by and say hello.

    Wanna write?

    Got that latest or greatest scoop? Perhaps you have some important news for the Debian community? Submit a news item!

    Or perhaps you've written a rather ground breaking insight into some aspect of Debian and you feel compelled to share it with others? Knock up a longer editorial article and send it to the editors.

    Sponsorship

    DP is sponsored by Xinit Systems and kieser.net.

    Domains paid for and hosted by uklinux.net.

    Buy your Debian merchandise at DebianShop.com.

    Who's Online

    There are currently, 97 guest(s) and 11 member(s) that are online.

    You are Anonymous user. You can register for free by clicking here.

      
    HOWTOs: Take the plunge, the water is nice and warm (upgrading to ext3)
    Posted on Saturday, March 30 @ 10:47:02 GMT

    File Systems
    For those who have not yet taken the plunge and upgraded from ext2 to ext3, now is as a good a time as any to experience the awe that is this particular journalled filesystem. Read more for full details of how to plunge into the tranquil waters of ext3.

    Ext3 as it suggests is a new, enhanced version of the ext2 filesystem, which is the most common Linux filesystem. It's fast, it's clean, the code has been tried and tested and most importantly it works.

    As drive capacity increases, the problem of the dreaded fsck becomes more and more irritating, as it can take a very long time to check filesystems the size of modern disks. Journalled filesystems relieve this problem -- they don't make your data more secure, but they can help reduce the problems of a fsck on power failure for example. All write operations are recorded in a 'Journal' on the disk, operations are written to the journal and only when they have successfully been written to the disk are the entries removed from the journal. If a power failure occurs, then the write operations which may have not been completed or only half successfully can be carried out when the journal is replayed on the machine at boot up. Thus removing the need for a fsck on boot up.

    So how do I use this great world saving innovation? First you need to have an up-to-date version of e2fsprogs. The versions in woody and sid are both happy to talk ext3. You will also need a kernel with ext3 support included, I recommend you include this into the kernel, not as a module. Kernel compilation is beyond the scope of this document, however I suggest you read my article on compiling kernels the Debian way. Follow these instructions using a relatively recent kernel source, e.g. 2.4.18 or 2.4.17.

    Now that you have kernel support, you need to create the journals on your ext2 partitions. The great thing about ext3 is that it's built on top of ext2 and you can mount a cleanly unmounted ext3 partition at ext2. Switching to ext3 is also non-destructive, however I do recommend you backup any very very important data (you are doing that anyway, aren't you!).

    To create the journal, use tune2fs -j <partition device>. If the partition is mounted then a .journal file is created. It is important you do not delete this or back it up. If it isn't mounted, the journal is hidden on the disk. You can use a seperate drive to store the journal but this is beyond the scope of this document.

    If you are creating partitions from scratch, you can use mkfs.ext2 -j <partition device>. You need to edit /etc/fstab to tell it to mount the partitions as ext3. They way I do this is to just change the "ext2" to "auto". This will make mount try all filesystems including ext3, followed by ext2.

    I hope you find this article encourages you to dive in and upgrade to ext3!

     

     

    spellchecked by Joy 🙂 PHP-Nuke sucks ass, it changes my &lt; and &gt; entities back to the characters every time...

     
    Related Links

  • ReiserFS (NAMESYS)
  • Ext3 (Stephen Tweedie)
  • JFS (IBM)
  • Linux XFS (SGI)
  • More about File Systems
  • News by rob

    Most read story about File Systems:
    Filesystems - Part I

  • "HOWTOs: Take the plunge, the water is nice and warm (upgrading to ext3)" | Login/Create Account | 29 comments
    Threshold


    The comments are owned by the poster. We aren't responsible for their content.

    Re: Take the plunge, the water is nice and warm (upgrading t (Score: 0)
    by Anonymous on Saturday, March 30 @ 12:02:40 GMT

    I've just recently taken the plunge into the unknown, ReiserFS being the only journalled fs I've tried before, I found that ext3 works quite well, the only downside is a slight performance decrease with ext3, I went from 37 - 38 mb/s to around 30 - 36 average during some hdparm benchmarks, but the occational power failures and X hangups makes it all worth it.

    I have btw finally tried the preempt patch, I patched 2.4.18 and to my surprise, I found it quite stable, the performance increase showed upon the first reboot and X session, and the overall feel is much smoother. However as I started filling the screen with mpeg clips, BeOS style, aswell as the glxgears demo and dragging files between nautilus windows, a bit overkill, yes. Doing a cold reboot after X decided to freeze, again showed ext3's strengths.

    [ No Comments Allowed for Anonymous, please register ]


    Re: Take the plunge, the water is nice and warm (upgrading to ext3) (Score: 2, Informative)
    by hildeb on Saturday, March 30 @ 14:01:08 GMT
    (User Info) http://www.arschkrebs.de/

    Note that ext3 doesn't protect you from the drive f*cking up.

    Most problems posted on the ext3-users list are due to bad hardware or hardware going bad.

    Keep that in mind.

    Also: Performance will most likely suffer a bit. Find out which mode (data=writeback or data=journal or the default data=ordered) is best for you.

    data=writeback:

    Journals only metadata. Like XFS, or ReiserFS

    data=journal:

    Journal metadata && userdata. Often slow, since all data is written to disk twice.

    data=ordered:

    Metadata is written to disk AFTER the userdata was written.

    [ No Comments Allowed for Anonymous, please register ]


    Re: Take the plunge, the water is nice and warm (upgrading to ext3) (Score: 1)
    by fan on Saturday, March 30 @ 17:49:15 GMT
    (User Info)

    Congratulations! Although i recently upgraded to ext3 and therefore this tutorial is useless for me, it just shows newbies and non-informed guys how to upgrade safely. Really a nice one!

    see you, fan

    [ No Comments Allowed for Anonymous, please register ]


    Re: Take the plunge, the water is nice and warm (upgrading t (Score: 0)
    by Anonymous on Sunday, March 31 @ 00:58:13 GMT

    Where's the tradeoff?

    If I'm not mistaken, a write operation is done to a logbook, then the application is informed: the write is succesfull en then the logbook is copied to the real location on the disc??

    Can somebody explain the technical basics to me?

    Thanks

    [ No Comments Allowed for Anonymous, please register ]


    Things of note: (Score: 2, Informative)
    by cef on Sunday, March 31 @ 06:05:48 BST
    (User Info)

    If you're using a recent version of e2fsprogs (ie: the latest ones in sid and woody):

    1. It will automatically remove the .journal files at the next boot, and allocate the blocks internally in the fs. This means that after a reboot the .journal file will dissapear - don't worry, this is correct behaviour. Just don't delete it yourself, let the FS tools take care of it for you.

    2. You can specify multiple types of FS's to try to mount for a particular entry in fstab, by seperating them with a comma. eg: 'ext3,ext2' in the fstype field is quite valid, and more reliable than auto, which if some sort of horrid disk errors occurs (such as XP deciding to 'own' your partitions), will prevent it trying to mount the disk as vfat or somesuch, and possibly corrupting the disk further.

    Also, if you boot your machine frequently, and don't have any issues, you might want to change the default fs check from 20 reboots to something higher. It can also be useful to stagger these if you have more than one partition so that you don't get them all colliding at once, checking all the partitions at the same boot. Don't disable it though, as the drive could still get physical errors that would go undetected, at least till you lost data!

    I've removed ext2 support from my normal boot kernel, and everything runs quite well. I also happily have the option of booting back to a pre-ext3 compiled kernel without any hassle. I'm exceptionally happy with ext3.

    Now all I'm waiting on is h-tree indexing for ext3. This will speed up directory indexing amazingly, and the preliminary implementations show real promise. It's also completely backward compatible with the current ext3/ext2 code, so there is no real upgrade issue. Bring it on!

    [ No Comments Allowed for Anonymous, please register ]


    Based on: PHP-Nuke

    All logos and trademarks in this site are property of their respective owner. The comments are property of their posters, all the rest © 2000 by Debian Planet

    You can syndicate our news using the file backend.php.