Page 2 of 2 FirstFirst 12
Results 11 to 13 of 13

Thread: ntfs-3g culprit of Ubutu's touted sluggishness (with intensive HDD I/O)?

  1. #11
    Join Date
    Apr 2007
    Location
    Mexico City, Mexico
    Posts
    899

    Default

    Quote Originally Posted by Svartalf View Post
    Why pick a consistently LOW performer when you can have the high performance for things overall and then have a few gotchas that bring it back down to the low performer's overall level?
    Well, for some reason (despite its performance, which for me has been rather good... for my needs that is) Ext3 is the best supported filesystem across distributions, not to mention is touted to be "Linux's native file system". I really do not care that much about a filesystem. All I want is decent enough access times to write and fetch my data, and my data to remain uncorrupted as much as possible (since I'm a physician, you can imagine this is of especial interest for me, I don't want any patient's records to be corrupted [which also prompts for rather frequent backups]).


    Quote Originally Posted by Svartalf View Post
    That has me a bit concerned to be honest since I've seen occasions where that this is the case. XFS isn't it either- while it's VERY fast, it only journals metadata and has issues on other odd edges as does JFS. I'm kind of keenly waiting to see what comes of btrfs at this point (Not anywhere near ready for primetime- but promises to be at least as good as reiser4 and possibly better than ZFS in many ways...). If Sun hadn't of insisted on the licensing they put on ZFS, this would be mostly a moot point, I think.
    I had been playing with the idea of converting my partitions to XFS, but as you point out, it doesn't implement a full journal, which kind of concerns me... However, AFAIK is the preferred FS for media content-creation as it has unsurpassed I/O performance with especially big files (movie raw data files?). I wish Sun did allow for a more flexible license on their ZFS (they know that by itself it is a GREAT asset ), and have been off the loop in the recent developments of other "better" FSes, so I don't know "what's out there". I know Ext4 debuted in one of the 2.6.2x kernels, but I do not know what exactly does have over Ext3, or its performance, etc. For the time being, I'm happy with what I've got, though.

  2. #12
    Join Date
    Jun 2006
    Posts
    3,046

    Default

    Quote Originally Posted by Thetargos View Post
    Well, for some reason (despite its performance, which for me has been rather good... for my needs that is) Ext3 is the best supported filesystem across distributions, not to mention is touted to be "Linux's native file system". I really do not care that much about a filesystem. All I want is decent enough access times to write and fetch my data, and my data to remain uncorrupted as much as possible (since I'm a physician, you can imagine this is of especial interest for me, I don't want any patient's records to be corrupted [which also prompts for rather frequent backups]).
    You should actually care about most of the piece-parts of your OS install. The corrupted records is of concern. Oh, and "best supported" doesn't mean anything other than people poring over the code and "fixing" things- it does mean brown paper bag over the head problems get found quickly, but it doesn't go that it's the best choice (Keep in mind- the thinking, even from the main kernel crowd, is that if btrfs comes in as good as it's still looking it'll replace everything else as the mainline filesystem... And it's got things that may leapfrog ZFS, which IS one of the best filesystem designs to date...).

    What does it matter if my FS is "robust" if I can't use the box for minutes at a time because I'm doing something disk-intensive... Is the answer jump to a new distribution? Not always. There's other...issues...that are not unlike this one with Ubuntu in Fedora, Mandriva, etc... Every one of them make dumb decisions from time to time.



    I had been playing with the idea of converting my partitions to XFS, but as you point out, it doesn't implement a full journal, which kind of concerns me... However, AFAIK is the preferred FS for media content-creation as it has unsurpassed I/O performance with especially big files (movie raw data files?). I wish Sun did allow for a more flexible license on their ZFS (they know that by itself it is a GREAT asset ), and have been off the loop in the recent developments of other "better" FSes, so I don't know "what's out there". I know Ext4 debuted in one of the 2.6.2x kernels, but I do not know what exactly does have over Ext3, or its performance, etc. For the time being, I'm happy with what I've got, though.
    Ext4 MIGHT be okay, but unfortunately, you can't boot with it right at the moment- however, it may have the same glitch that Ext3 currently has. XFS and JFS don't do anything other than metadata journalling, but in and of itself with a UPS, you should mostly be fine with them- and you CAN boot with either of those or Reiserfs (though we know what gives with this...I probably need to move to XFS or JFS on my home dir here at work...). Btrfs is in development and is expected to be mostly ready for prime-time in 6 or so months. There's issues with it that are currently unresolved (as in not fully implemented yet...) so I wouldn't use it for anything important. ZFS is available as a fuse filesystem and could be a choice for some things.
    Last edited by Svartalf; 11-14-2008 at 12:23 PM.

  3. #13
    Join Date
    Sep 2008
    Posts
    130

    Default

    Forgive me if I'm stating the obvious here, but couldn't it just be that the kernel is not tuned for desktop? Last time I looked at the Ubuntu Kernel sources the I/O latency settings were set to what was essentially the server option

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •