The 4 raid5 drives here all only hold one partition that fills the whole
drive. No lilo, nothing, since I boot from an additional disk (on a
different controller).
fdisk shows:
/dev/hde1: begin 1, start 1, end 32760, blocks 16511008, type 83, linux
native
seems all ok.
My little problem her stems most likely from my somewhat nasty badblocks
overwrite. Also possibly from the combination of 2.0.36 + raid + libc5.
However, on a different machine, I also experienced some (not raid
related) problems with several partitions on the same disk. Potentially
caused by dos fdisk, with which I usually install loadlin dos boot
partitions. With the large hds the partitioning seems to get somewhat
tricky.
The "begin" and "start" issue is pretty strange. The suse linux
partitioner just shows "from" and "to".
---
Thilo Herrmann, Rival Network GmbH
Phone: +49 89 82087-0, email: [EMAIL PROTECTED]
---
-----Original Message-----
From: Tom Brown <[EMAIL PROTECTED]>
To: Thilo Herrmann <[EMAIL PROTECTED]>
Cc: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date: Mittwoch, 16. Dezember 1998 22:46
Subject: Re: 98-12-14/15, raid5, some test results, some non-fatal
problems
>
>>
>> /var/log/messages after badblocks -w /dev/hde & raidstart /dev/md0
>>
>> Dec 16 19:20:26 fileserver kernel: (read) hde1's sb offset: 16510912
>> [events: aaaaaaaa]
>> Dec 16 19:20:26 fileserver kernel: md: invalid raid superblock magic
on
>> hde1
>> Dec 16 19:20:26 fileserver kernel: md: hde1 has invalid sb, not
>> importing!
>> Dec 16 19:20:26 fileserver kernel: could not import hde1!
>> Dec 16 19:20:26 fileserver kernel: autostart hde1 failed!
>> Dec 16 19:20:26 fileserver kernel: huh12?
>> Dec 16 19:21:43 fileserver kernel: md: can not import hde1, has
active
>> inodes!
>
>Hhmm... you're running into the same problem I posted last night, which
>seemed to be from the wierd things that happen when you use big IDE
drives
>and partitions that start > 1024 cylinders... I still haven't gone
back
>and finished dealing with that...
>
>you might want to post the output from fdisk -l ... no, that won't list
>drives past hdd, you'll have to print each partition table... and let
us
>know if anything else was using partitions on hde...
>
>While I'm here... can anyone point me to a resource that describes
what's
>happening when the BEGIN and START cylinders are not equal?
>
>----------------------------------------------------------------------
>[EMAIL PROTECTED] | The thing always happens that you really
believe in;
>http://BareMetal.com/ | and the belief in a thing makes it happen.
>web hosting since '95 | - Frank Lloyd Wright
>
>
