Re: How should this issue be handled, should it be taken to the Technical Ctte?
#include <hallo.h>
* Herbert Xu [Mon, Feb 17 2003, 07:36:21AM]:
> > And when did this happen? How long is LVM already beeing used?
>
> I've been using LVM before that bug was filed. However, as I said,
Fine, me too.
> > Fine. What about some indication about this progress in the BTS?
>
> Doing so with this message.
Nice, thank you.
> > And what about that dynamic raid autodetection, IIRC I have already
> > writen that you method (raidtools and static config) is bad. In the
> > meantime I dicussed the problem with another DD and the best idea was
> > patching mdadm to work directly with major/minor numbers.
>
> Well I must say I missed that particular message since I was overseas
> at the time. The issue there is we must preserve the RAID device
> ordering or chaos will ensue. If you've got a patch that does that,
> I'm sure it will speed up the resolution of this issue.
Okay, I attach the patch I created I while ago. It uses mdrun from mdadm
package (Sid) to check all available partitions, detect raid volumes on
them and activate raid sets. Unfortunately, it won't work with the
current version of initrd-tools and a kernel without devfs, since you
create device nodes in tmpfs. It may work if we create some provisoric
/dev directory. I have an idea how to modify mdrun to work this way, but
I do not have much time in the next days to do it. You will hear from
me.
Gruss/Regards,
Eduard.
--
Nearly all men can stand adversity, but if you
want to test a man's character, give him power.
-- Abraham Lincoln
Reply to: