Messages in this thread | | | Date | Wed, 6 Dec 2000 07:12:50 -0800 (PST) | From | Tom Murphy <> | Subject | IDE fs corruption in 2.4.0-test11? |
| |
My system running 2.4.0-test11 recently trashed itself and I was wondering what steps I should take to ensure UDMA is running OK and not trashing itself.
Are there options in hdparm that could conflict with 2.4's handling of UDMA that would cause all writable partitions on the drive to self-destruct? This is what happened to me last weekend except for /usr, which was mounted read-only at the time.
2.4.0-test11 spontaneously just rebooted itself for no reason, so I decided booting into 2.2.17 would be a good idea to check to make sure everything was fine. 2.2.17 fsck'd all the partitions and was happy and went on.
I then rebooted back into 2.4.0-test11 (or tried to) and LILO greeted me with a lovely "LIL-".
I booted from rescue disks and attempted to fsck the partitions. I had to specify the superblock and eventually got a huge string of "Illegal blocks in inodes". After fscking "/", I noticed /etc was gone. Same thing happened in /var and /home with basically no data being fixed or recovered. I had to reformat everything.
Which options in hdparm are considered dangerous to use in 2.4? I had 32-bit I/O set (-c 1) , DMA set on (-d 1), multiple sector I/O (-m 16). The drive in question is a 45 gigabyte Western Digital Caviar drive.
I heard the multiple sector I/O may be bad for the drives.. is this true?
Thanks in advance,
Tom
ps. Please reply directly to my e-mail address as I am not subscribed to linux-kernel.
__________________________________________________ Do You Yahoo!? Yahoo! Shopping - Thousands of Stores. Millions of Products. http://shopping.yahoo.com/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
| |