On Thursday August 31, [EMAIL PROTECTED] wrote:
> I am seeing the following messages on 2 raids set up on 2.4.0-test4,
> after I upgraded to 2.4.0-test7. I have
>
> CONFIG_MD_BOOT=y
> CONFIG_AUTODETECT_RAID=y
>
> and appended "md=..." params to lilo (is this required for autodetect?)
>
> This is my system:
>
> hanagumi:~# raidstart --version
> raidstart v0.3d compiled for md raidtools-0.90
> hanagumi:~# uname -a
> Linux hanagumi 2.4.0-test7 #1 SMP Wed Aug 30 21:11:48 JST 2000 i686 unknown
> hanagumi:~# gcc -v
> Reading specs from /usr/lib/gcc-lib/i386-linux/2.95.2/specs
> gcc version 2.95.2 20000220 (Debian GNU/Linux)
> hanagumi:~#
>
> On startup I see the following messages, and the raids fails to mount.
> /proc/mdstat shows the raids are started and normal.
>
> If I "raidstop/raidstart" the raids they seem to work fine, but I did not experiment
>too much on them.
>
> Can somebody look into it? My kernel config is included.
>
> Stephen
>
> Aug 30 21:30:44 hanagumi kernel:
> Aug 30 21:30:44 hanagumi kernel: md1 stopped.
> Aug 30 21:30:44 hanagumi kernel: md: bug in file md.c, line 164
> Aug 30 21:30:44 hanagumi kernel:
The problem is that you are using both MD_BOOT and AUTODETECT to start
the raid arrays, and the code isn't coping.
With 2.4.0, you have a choice for configuring raid at boot time. You
can use autodetect by setting partition types, or use md= arguments to
the kernel. You can even use both for different arrays, but not both
for the same array.
For following patch should make it cope better if you try to do both.
NeilBrown
--- md.c 2000/08/31 23:07:29 1.1
+++ md.c 2000/08/31 23:08:57
@@ -3745,6 +3745,10 @@
if (!(md_setup_args.set & (1 << minor)))
continue;
printk("md: Loading md%d.\n", minor);
+ if (mddev_map[minor].mddev) {
+ printk(".. md%d already autodetected - use
+raid=noautodetect\n");
+ continue;
+ }
mddev = alloc_mddev(MKDEV(MD_MAJOR,minor));
if (md_setup_args.pers[minor]) {
/* non-persistent */
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to [EMAIL PROTECTED]
