Alvin Oga | 1 Jul 01:50 2002

Re: Can't boot from /dev/md0 (RAID-1) (more info)


hi ya

how was the floppy made ??
	- does it have a initrd file ??
	- initrd is notlisted in your lilo.conf.hda
	- is the kernel on the bootable floppy the same as that
	referenced in lilo.conf ?

i'd take out all those explicit defines and see if it helps

simplified(?) lilo.conf
	boot=/dev/md0
		- not sure about boot= for /dev/hda vs /dev/md0
	map=/boot/map
	install=/boot/boot.b
	prompt
	timeout=50
	lbad32

	image=/boot/vmlinuz	   
	   root=/dev/md0
	   label=raid
	   read-only

-- you'd also want to use /dev/hda  and /dev/hdc ...

c ya
alvin

(Continue reading)

Roy Sigurd Karlsbakk | 1 Jul 12:07 2002
Picon

Re: Can't boot from /dev/md0 (RAID-1)

On Sunday 30 June 2002 23:46, Michael wrote:
> > I hope this is not OT - didn't find any LILO mailing list. after
> > trying virtually everything - can anyone help me with a tip?
>
> The problem you're having is not with raid but with LILO in general.
> I don't remember the specifics, but I've experienced the same thing
> and it was an error in how I used lilo or the version of the boot
> record -- something strange like that. Raid is not part of the
> problem.
>
> Read the docs that come with the LILO distribution in the source. The
> man page does not have as much info. You may also be suffering from
> differing version of the MBR -- a major source of pain and suffering.

I've read the fscking manual - all I can find - for both grub and lilo, and 
I've tried all possible configurations I can think of.

point is - I beleive there's a BIOS/linux misunderstanding somewhere

roy

--

-- 
Roy Sigurd Karlsbakk, Datavaktmester

Computers are like air conditioners.
They stop working when you open Windows.

-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo <at> vger.kernel.org
(Continue reading)

Roy Sigurd Karlsbakk | 1 Jul 12:17 2002
Picon

Re: Can't boot from /dev/md0 (RAID-1) (more info)

Please listen

I've tried all this

see the dmesg output

On Monday 01 July 2002 01:50, Alvin Oga wrote:
> hi ya
>
> how was the floppy made ??
> 	- does it have a initrd file ??
> 	- initrd is notlisted in your lilo.conf.hda
> 	- is the kernel on the bootable floppy the same as that
> 	referenced in lilo.conf ?
>
> i'd take out all those explicit defines and see if it helps
>
> simplified(?) lilo.conf
> 	boot=/dev/md0
> 		- not sure about boot= for /dev/hda vs /dev/md0
> 	map=/boot/map
> 	install=/boot/boot.b
> 	prompt
> 	timeout=50
> 	lbad32
>
> 	image=/boot/vmlinuz
> 	   root=/dev/md0
> 	   label=raid
> 	   read-only
(Continue reading)

Jakob Oestergaard | 1 Jul 13:21 2002
Picon

Re: Can't boot from /dev/md0 (RAID-1)

On Mon, Jul 01, 2002 at 12:07:07PM +0200, Roy Sigurd Karlsbakk wrote:
...
> 
> I've read the fscking manual - all I can find - for both grub and lilo, and 
> I've tried all possible configurations I can think of.
> 
> point is - I beleive there's a BIOS/linux misunderstanding somewhere

Did you try the   raid-extra-boot  option for LILO ?  I really don't
know if this is the problem you are seeing, but it's worth a shot I
think.

--

-- 
................................................................
:   jakob <at> unthought.net   : And I see the elder races,         :
:.........................: putrid forms of man                :
:   Jakob Ƙstergaard      : See him rise and claim the earth,  :
:        OZ9ABN           : his downfall is at hand.           :
:.........................:............{Konkhra}...............:
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo <at> vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Flemming Frandsen | 1 Jul 13:29 2002
Picon

Non-equal sized partitions vs RAID 1

I have two disks in my box, both are MAXTOR 6L080L4 (80GB IDE) and 
physically the same disks, both have 3 partitions, one for /boot one for 
swap and one huge one for the rest for the mirrored data.

The mirror works as expected, but somewhere something went slightly 
wrong because one disks geometry is:
physical     155114/16/63
logical      155114/16/63

The other disk is:
physical     155114/16/63
logical      9732/255/63

Now, this has the annoying sideeffect of making it impossible for me to 
partition them exactly the same way, so I ended up creating the mirror 
on disk 1 (which has a slightly smaller partition for mirroring than 
disk 2) and then adding disk 2 later.

This will get me in trouble (I tested it) when disk 1 fails and I need 
to replace it, because then I have to make that partition slightly 
larger than the disk 2 partition so it can be added to the mirror.

Now for the questions:
* Is there a way to change the logical geometry of disk 2, so I can 
partition correctly (well if there is then this post is OT)?

* Is there a way to force the array to keep being the same size, even 
after the smaller partition has failed?

* Why does the BIOS think that it's cool to translate similar disks 
(Continue reading)

Danilo Godec | 1 Jul 13:54 2002
Picon

Re: Non-equal sized partitions vs RAID 1

On Mon, 1 Jul 2002, Flemming Frandsen wrote:

> Now for the questions:
> * Is there a way to change the logical geometry of disk 2, so I can
> partition correctly (well if there is then this post is OT)?

You can use 'hdx=c,h,s' kernel parameter on boot. You can also use fdisk's
expert menu to change the geometry while creating the partitions. I don't
know if the later alone is enough, I prefer to use the kernel parameter
and only use the fdisk trick to skip one reboot.

Remember to replace 'hdx' with your actual device name... :)

> * Is there a way to force the array to keep being the same size, even
> after the smaller partition has failed?

After the array is created, it doesn't (shouldn't) change.

> * Why does the BIOS think that it's cool to translate similar disks
> differently, are BIOS writers crackheads or was I just unlucky?

All BIOSes seem to do that: Disks, connected to the primary IDE interface
show 'correct' geometry, while disks, connected to the secondary interface
don't. I hoped it would go away with recent bioses, but I'm affraid I'll
just have to get used to it... :/

   D.

--

-- 
___________________________________________________________________
(Continue reading)

Cal Webster | 1 Jul 18:58 2002

md: badblocks(pid 1216) used obsolete MD ioctl

Hello list:

I'm getting errors when formatting and managing RAID devices. Below is the
sequence of events leading up to the errors. I had recently upgraded our
RAID5 array from 6 to 9 disks (see below for system profile and raid
config). To do this, I backed up the contents of the RAID, formatted the new
drives, and re-created it with a new raidtab. There were no errors on any
RAID devices prior to the upgrade.

Please note that the same sector is shown in each of the two I/O errors, so
there may be a real error. However, why can't I get "fsck" to work around
it? Why did the sector error not show up before? Also, why did I only get
data errors in the filesystem check? Isn't the "badblock" test supposed to
catch these bad blocks? Could the "obsolete MD ioctl" errors be keeping
"fsck" from correcting the errors?

Thanks!

--Cal Webster
Network Manager
NAWCTSD ISEO CPNC

############################
# Begin Sequence of Events #
############################

1. Upgraded from 6 to 9 drives - no apparent errors during creation or
restoration of RAID5 device.

2. After restoring data, /dev/sdc1 got kicked from array.
(Continue reading)

Cal Webster | 1 Jul 21:43 2002

RE: md: badblocks(pid 1216) used obsolete MD ioctl

> -----Original Message-----
> From: Maurice Hilarius [mailto:maurice <at> harddata.com]
> Sent: Monday, July 01, 2002 10:41 AM
> To: cwebster <at> ec.rr.com
> Subject: Re: md: badblocks(pid 1216) used obsolete MD ioctl
>

Thank you for the feedback Maurice.

> You have SCSI bus problems. Likely related to termination.
> 1) check if you have active, powered termination, of the correct type. A
> U160 term will generally work, even with lower SCSI variants, as long as
> they are LVD of some type.

Forgive me, but I don't believe this is a SCSI bus problem. If the problem
moved around or I got different errors each time, I might suspect the bus.
As it is, the same sector on the same drive is called out each time. These
drives are mounted in an external SCSI drive enclosure which is terminated
with the appropriate active termination at the opposite end of the bus from
the connection to the host computer.

> 2) Check if the devices are jumpered to provide termination power to the
> SCSI bus. Make it so if possible, on ALL devices.

All the new drives are Seagate Barracuda ST318438LW. Unless I am mistaken,
they are ready to install in their factory default configuration.
Termination power is supplied by the bus for the external active terminator.
None of the nine drives should be terminated or jumpered to supply
termination power.

(Continue reading)

Thomas Hicks | 1 Jul 23:04 2002

Mdadm issue

This is the first time I am setting up raid using the mdadm tool. This
is what I am trying to do. 

The sever is as follows: 2 drives at 10gig
                         2 drives at 80gig
hda2 is /
hdd1 is my /home

I have the other drives in ad ready and would like to set up the 2 10
gig's as a mirror and the 2 80 gig as raid0

At the shell I run this command 

mdadm --create /dev/md0  --level=1 --raid-disks=2 /dev/hda2 /dev/hdb1
and get the following output:

mdadm: /dev/hda2 appears to contain an ext2fs file system
    size=7269412K  mtime=Mon Jul  1 11:56:16 2002
mdadm: /dev/hdb1 appears to contain an ext2fs file system
    size=7333640K  mtime=Mon Jul  1 11:56:29 2002
Continue creating array? y
mdadm: SET_ARRAY_INFO failed for /dev/md0: File exists

Could someone please help me with this problem. I am having a really
hard time finding any info on the web about this. 

-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo <at> vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
(Continue reading)

Cal Webster | 2 Jul 21:29 2002

RE: md: badblocks(pid 1216) used obsolete MD ioctl

> -----Original Message-----
> From: Maurice Hilarius [mailto:maurice <at> harddata.com]
> Sent: Monday, July 01, 2002 2:26 PM
> To: cwebster <at> ec.rr.com
> Subject: RE: md: badblocks(pid 1216) used obsolete MD ioctl
>
> I missed that. Always the same drive and sectors?

That's right: single drive, one bad sector.

> It might be worthwhile to run a SCSI (low level) format on this drive, so
> that any bad sectors get marked bad and are not used.

The tools at my disposal relevant to this task are fdisk, mke2fs, e2fsck,
and badblocks. I don't believe that the "surface analysis" done by most SCSI
BIOS utilities on Intel machines does much more than these tools.

> It never hurts to have device close to the terminator ALSO providing term
> power. By the time the current gets from the host, all the way to
> the other
> end of the cabling, especially with an external cabinet, the voltage may
> drop quite a bit.
> It does no harm to enable it on the drives as well, and makes sure the
> attenuation is not a problem.

I appreciate what you're saying, especially since I did not indicate the
proximity of the RAID array to the host computer. However, the length of
wire between this on-board interface connector and the last device in the
external array is less than 8 feet total, so I would expect zero benefit. To
the contrary, in our situation all it would do is make the drives run hotter
(Continue reading)


Gmane