Broken Ext3 or dead HDD ???
FedoraForum.org - Fedora Support Forums and Community
Results 1 to 8 of 8
  1. #1
    Join Date
    Jun 2007
    Posts
    97
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)

    Broken Ext3 or dead HDD ???

    I have a strange problem:

    When downloading and checking the sha1 sum of big files I never get consistent results.

    All of the Fedora iso images have a publicly known sha1 sum. When I download them and verify the sum I never get it right. If I copy the downloaded file to other dir and check the sha1 of both the original download and the copied file I get 2 different sums. If I restart the computer and recheck the downloaded and the copied file I get another different sha1sums inconsistent with the sums before the restart.

    I tried this with different big iso images and always get different results.

    Is my HDD dead or the ext3 journal has some problems. My HDD is a 80GB Sata Hitachi (notebook) 1 year old. It is supported with sata_sil module and lib_ata. Maybe they've kind of killed it (spin ups and downs).

    What should i do to check this? The check tools on the official hitachi site won't run on my notebook. Any Ideas?

  2. #2
    Join Date
    Jun 2007
    Posts
    97
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    Anyone? Anyideas?

    I noticed that even using chown and chmod on one the same file gives me different sha1 after the operation. It's crazy! I don't think that chown and chmod change the file in anyway. This doesn't happen all the time.

    This makes me think it is something connected to the ext3 filesystem.

  3. #3
    Join Date
    Oct 2008
    Posts
    299
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    Quote Originally Posted by pepe123
    Anyone? Anyideas?

    I noticed that even using chown and chmod on one the same file gives me different sha1 after the operation. It's crazy! I don't think that chown and chmod change the file in anyway. This doesn't happen all the time.

    This makes me think it is something connected to the ext3 filesystem.
    They change the file attributes, but I'm not sure if that's IN the file, it could be, but don't ask me.


    Nevertheless, I usually save things on my NTFS partition to make sure it's consistent without the extra attributes.

    Anyways, if it REALLY is corrupt, I used to have the same problem, and I ended up having to switch ISP's because they had some lame thing that corrupted most of my downloads.
    Intel Core 2 Quad Q8400 CPU (2.66 GHz)
    3gb RAM
    500GB HD
    Debian Lenny x86_64 & Windows 7 Ultimate x64
    NVIDIA GeForce 9800 GTX+ Superclocked - Fattypack

    Second(Server) PC:
    512mb RAM
    80gb TiVO HD(Lololol, took it from TiVO, old one broke)
    PC-BSD 7.1.1 x86
    Ossom router decked out with DD-WRT as a bridge providing net
    Sign the Online Petition to bring Crysis 2 to Linux/Mac!

  4. #4
    Join Date
    Sep 2006
    Location
    Connellsville, PA, USA
    Posts
    11,308
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    Use smartmontools (smartctl -H /dev/sd[a-z]) to assess HDD health. But from your description, it's more likely bad RAM. Run memtest86+ overnight (at least 8 hrs.) on the RAM.

    V

  5. #5
    Join Date
    Jun 2007
    Posts
    97
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)

    It's the EXT3 that causes me the problem :(

    Thank you for the replies guys,

    Well it is the ext3 journal that causes me the problem.

    I run fsck and e2fsck form rescuecd and they corrected some errors on the /home and /root partitions. after this I tried again copying and moving (even packing and repacking) the big files (they are over 6 GB! each) and got no errors this time. chmod and chroot didn't show inconsistencies either.

    I tried memtest86+ - no errors. it run for 4 hours and still ok. About the ISP - i can correctly download the files on win mashine, so it's not the problem.

    I'll keep on eye on the disk for a couple of days and report back.

    I don't believe that the 1year old HDD is dead. it is possible but i don't want to believe it right now.

  6. #6
    Join Date
    Jun 2007
    Posts
    97
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    This is really strange. smartmodtools didn't show any error. the short ang long tests completed without errors. But something isn't still in ordnung.

    when copying and mv a file between directories it shows different sha1 sums every time . This is impossible!!! I didn't do anything with the file. just issued sha1sum command 2 consequtive times on one and the same file and got 2 different results!!!

    Maybe I should file a bug report bug against what? ext3 maybe?

  7. #7
    Join Date
    Sep 2006
    Location
    Connellsville, PA, USA
    Posts
    11,308
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    I tried again copying and moving (even packing and repacking) the big files (they are over 6 GB! each) and got no errors this time. chmod and chroot didn't show inconsistencies either.
    I vaguely recall some issues with very large file sizes. You may want to investigate from that angle.

    And run memtest86+ for at least 8 hours.

    V

  8. #8
    Join Date
    Jun 2007
    Posts
    97
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    It was faulty ram modules. I tested them and one got errors at some address. I changed them and everything is fine now.

Similar Threads

  1. Replies: 7
    Last Post: 7th January 2014, 07:12 PM
  2. Replies: 1
    Last Post: 6th April 2009, 10:35 AM
  3. yum is DEAD
    By thomwickes in forum Using Fedora
    Replies: 1
    Last Post: 29th September 2007, 12:48 PM
  4. Can't read EXT3 fs, completely dead!
    By MonkeyMan in forum Using Fedora
    Replies: 2
    Last Post: 16th July 2006, 06:46 PM

Tags for this Thread

Posting Permissions

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