Messages in this thread | | | From | Russell Coker <> | Subject | 2.4.0-test2-ac2 with RAID lockup | Date | Sun, 2 Jul 2000 23:48:55 +0200 |
| |
I am using RAID-1 with two 46G IBM ATA-66 hard drives, the RAID-1 is on /dev/hda4 and /dev/hda4 and is 43510604 blocks in size.
Here's the fdisk output from /dev/hdc, /dev/hda is exactly the same (I used dd to copy the partition table).
Disk /dev/hdc: 255 heads, 63 sectors, 5606 cylinders Units = cylinders of 16065 * 512 bytes
Device Boot Start End Blocks Id System /dev/hdc1 1 33 265041 82 Linux swap /dev/hdc2 34 36 24097+ 83 Linux /dev/hdc3 37 189 1228972+ 83 Linux /dev/hdc4 190 5606 43512052+ 83 Linux
Here is my current /etc/raidtab: raiddev /dev/md0 raid-level 1 nr-raid-disks 2 persistent-superblock 1 chunk-size 4k device /dev/ide/host0/bus0/target0/lun0/part4 raid-disk 0 device /dev/ide/host0/bus1/target0/lun0/part4 raid-disk 1
Here is my df output: Filesystem 1k-blocks Used Available Use% Mounted on /dev/ide/host0/bus0/target0/lun0/part3 1228916 952728 276188 78% / /dev/ide/host0/bus0/target0/lun0/part2 23333 4172 17957 19% /boot /dev/md/0 43510604 557652 42952952 1% /mnt
I am migrating to devfs as you will notice. I started playing with devfs AFTER the first time RAID locked up on me so I know it's not caused by devfs. I have had RAID lockup with and without devfs.
The first time it locked up I could run ps/top/etc but any attempt to sync etc would hang (D state). The second time it was more serious. The command: ls -al /proc/247 would go D state, so ps, top, etc would all hang. Pid 247 was named.
At the time of the hang the only process using the RAID device was a bonnie++ test run on a ReiserFS file system. I doubt that ReiserFS or the drive hardware was at fault because I spent an entire week running various bonnie++ tests on /dev/hda4 with ReiserFS and ext2 without any problems at all. All the evidence I have suggests that ReiserFS is not at fault, however I have CC'd the ReiserFS list (just in case) and I am continuing to run tests.
The bonnie++ command was: bonnie++ -s 512
It seems that running the bonnie++ command on it's own does not cause a hang, it's only when I redirect stdout and stderr to files on the same file system that problems occur. I've just run another test and had it hang badly enough that I couldn't login via telnet or ssh (the machine has no monitor).
This time it got through the first run but on the second run (with parameters "-s 1024 -n 60") it hung.
The machine in question is a test machine. So if anyone has any experimental patches etc they want me to run then please let me know.
Russell Coker
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |