greg.martin at ddos.com
2005-Feb-08 20:38 UTC
Ext3 Journal corruption on hitachi deskstars
I recently came across an enormous cluster of x86 clone machines running fedora core 1 (2.4.24) which have typically all intel or amd have VIA IDE chipsets. They frequently experience corrupted journals rendering the ext3 partition in read-only mode. More important than recovering the filesystem, I am interested in finding the root of the problem. The common hardware that all of these machines share is a Hitachi deskstar drive (New IBM drive)... Searching through the mailing list and google I have seen references to problems with this particular drive but for the life of me and (the broken redhat search tool) I cannot pin down specific information on the cause of these failures. A bug within the IDE bus driver and the harddrive, seems most sensible. Please help me find the right place to look. Thanks in advance, -Greg -- Greg Martin Sr. Information Security Engineer Melior Inc. www.ddos.com
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 greg.martin at ddos.com schrieb:> They frequently experience corrupted journals rendering the ext3 > partition in read-only mode. More important than recovering the > filesystem, I am interested in finding the root of the problem. > The common hardware that all of these machines share is a Hitachi > deskstar drive (New IBM drive)...maybe you can elaborate a bit more on the "corrupted journals": what does "fsck" say, what's in the kernel log (during mount). if we know the symptoms, perhaps someone can find the root of the problem... Christian. - -- BOFH excuse #161: monitor VLF leakage -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.5 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFCCVasC/PVm5+NVoYRAlDHAKDhp8Ep0S2J9NO8xfewnOMKdo1TEgCgsvhW 6r5zEZIJbznNz7+qd2X67VI=GavO -----END PGP SIGNATURE-----
Hi, On Tue, 2005-02-08 at 20:38, greg.martin at ddos.com wrote:> I recently came across an enormous cluster of x86 clone machines running > fedora core 1 (2.4.24) which have typically all intel or amd have VIA IDE > chipsets. > > They frequently experience corrupted journals rendering the ext3 partition > in read-only mode.When ext3 first takes a filesystem offline/readonly, it always logs an initial message explaining why it's doing so. You'll need to capture that log message. Obviously, if it's the filesystem containing /var/ that went readonly, then /var/log/messages won't contain the information; but "dmesg" may, and you can set up serial or network console to capture it. --Stephen