Stumped adding new disk.

Stumped adding new disk.

Post by Mulliken » Sat, 29 Jul 2000 04:00:00



I'm stumped, and so is the guy at EMC (formerly Data General).

I'm trying to add some new disk to our Ultra Enterprise 3000 server,
running Solaris 2.6.  I have attached a Clariion RAID subsystem with 4
luns configured, two on each of two targets.  There was already one disk
enclosure on the SCSI II F/W interface, and I just daisy chained off of
that enclosure, moving the terminator to the end of the string.

The original enclosure housed one disk, c2t1d0, and it has been in use
for some time.  The new subsystem adds 4 new luns, which are t2d0, t2d1,
t3d2, and t3d3.  I edited the /kernel/drv/sd.conf file, to add t2d0 thru
t2d7 and t3d0 thru t3d7, and then did a reboot -- -r.

After reboot, a dmesg shows four messages, one for each of the new luns,
complaining about "corrupt label  -  wrong magic number".

The kicker is that when I run the format command, the four new luns
don't show up in the list, so I can't label them.  Additionally, they
don't show up as entries in /dev/rdsk.

Any ideas out there?  I'll try anything.  All the guy at EMC support
could suggest was that I restore the original 'default' sd.conf file and
do a reconfiguration reboot to see if lun 0 shows up for each of the new
targets added.  Anybody know where the original template for this file
might be found in Solaris 2.6?

You can email me, or reply to this, whichever you prefer.


Much appreciation will be comming your way, if you can throw me a rope
on this one.

Cheers...   Jeff Mulliken

 
 
 

Stumped adding new disk.

Post by Doug O'Lear » Sat, 29 Jul 2000 04:00:00


[This followup was posted to comp.unix.solaris and a copy was sent to the
cited author.]

Hey;

There's a command that you can run that will rebuild the device tree;
however, for the life of me, I can't remember what it is.  It's not much
more than a reconfig reboot though, so that would be my first
suggestion...

[[ few seconds and  a bit of research later ]]

Ahh, there they are:

Research these commands before executing them!!!

1.      "cd /devices; drvconfig -r" should rebuild your physical device
tree.  Based on what you're seeing in dmesg, doesn't sound like you
should have to do that.

2.      "disks -r" should rebuild the /dev/dsk & /dev/rdsk links.  
Following that, the format command should work.

Both these commands are run automatically on a reconfig reboot.  Seems
like that should be your first step.

HTH;

Doug
--
-------------------
Douglas K. O'Leary
Senior System Admin


 
 
 

Stumped adding new disk.

Post by loriwilliam » Sat, 29 Jul 2000 04:00:00


Try running the drvconfig and/or the disk command. If that doesn't work try
the
boot -r at the ok prom."Ignore that if that is what the reboot --r means.."


> I'm stumped, and so is the guy at EMC (formerly Data General).

> I'm trying to add some new disk to our Ultra Enterprise 3000 server,
> running Solaris 2.6.  I have attached a Clariion RAID subsystem with 4
> luns configured, two on each of two targets.  There was already one disk
> enclosure on the SCSI II F/W interface, and I just daisy chained off of
> that enclosure, moving the terminator to the end of the string.

> The original enclosure housed one disk, c2t1d0, and it has been in use
> for some time.  The new subsystem adds 4 new luns, which are t2d0, t2d1,
> t3d2, and t3d3.  I edited the /kernel/drv/sd.conf file, to add t2d0 thru
> t2d7 and t3d0 thru t3d7, and then did a reboot -- -r.

> After reboot, a dmesg shows four messages, one for each of the new luns,
> complaining about "corrupt label  -  wrong magic number".

> The kicker is that when I run the format command, the four new luns
> don't show up in the list, so I can't label them.  Additionally, they
> don't show up as entries in /dev/rdsk.

> Any ideas out there?  I'll try anything.  All the guy at EMC support
> could suggest was that I restore the original 'default' sd.conf file and
> do a reconfiguration reboot to see if lun 0 shows up for each of the new
> targets added.  Anybody know where the original template for this file
> might be found in Solaris 2.6?

> You can email me, or reply to this, whichever you prefer.


> Much appreciation will be comming your way, if you can throw me a rope
> on this one.

> Cheers...   Jeff Mulliken

 
 
 

Stumped adding new disk.

Post by Doug O'Lear » Sun, 30 Jul 2000 04:00:00


Quote:> Seems like that should be your first step.

And was.  It'd be nice if I read the original message all the way
through, hey?  Sorry 'bout that.

Doug
--
-------------------
Douglas K. O'Leary
Senior System Admin

 
 
 

Stumped adding new disk.

Post by Arun Gurjal » Mon, 07 Aug 2000 04:00:00


You must urn "/etc/init.d/drvconfig" but there a few things you must do
before you run drvconfig.  Here is the list:
1.  First check and see if the add-on disk is recognized.  At the OK prompt
run probe-scsi-all.  If it is being recognized then you are OK. The system
sees it.
2.  Now boot up in the regular fashion.
3.  At the prompt (Bourne shell) you need to set an environment variable and
this how you set it:
    _INIT_RECONFIG=TRUE;export _INIT_RECONFIG
4.  Now change directory to /etc/init.d and run ./drvconfig.
5.  After drvconfig run "./devlinks".  This makes the character specific
file in /dev for you to recognize the drive in the system.
Hope this helps.
Arun


> I'm stumped, and so is the guy at EMC (formerly Data General).

> I'm trying to add some new disk to our Ultra Enterprise 3000 server,
> running Solaris 2.6.  I have attached a Clariion RAID subsystem with 4
> luns configured, two on each of two targets.  There was already one disk
> enclosure on the SCSI II F/W interface, and I just daisy chained off of
> that enclosure, moving the terminator to the end of the string.

> The original enclosure housed one disk, c2t1d0, and it has been in use
> for some time.  The new subsystem adds 4 new luns, which are t2d0, t2d1,
> t3d2, and t3d3.  I edited the /kernel/drv/sd.conf file, to add t2d0 thru
> t2d7 and t3d0 thru t3d7, and then did a reboot -- -r.

> After reboot, a dmesg shows four messages, one for each of the new luns,
> complaining about "corrupt label  -  wrong magic number".

> The kicker is that when I run the format command, the four new luns
> don't show up in the list, so I can't label them.  Additionally, they
> don't show up as entries in /dev/rdsk.

> Any ideas out there?  I'll try anything.  All the guy at EMC support
> could suggest was that I restore the original 'default' sd.conf file and
> do a reconfiguration reboot to see if lun 0 shows up for each of the new
> targets added.  Anybody know where the original template for this file
> might be found in Solaris 2.6?

> You can email me, or reply to this, whichever you prefer.


> Much appreciation will be comming your way, if you can throw me a rope
> on this one.

> Cheers...   Jeff Mulliken

 
 
 

1. Adding new quorum device, unable to do a new disks probing...

Hi, I've configured new shared disks for my cluster's nodes (Sun Cluster
2.2). Each node see these devices through 'format' but I can't see them
whith 'finddevices' probe command in order to do a new quorum device
assignment. So I can do it, because neither "scconf <mycluster> -q <node1>
<node2>" command don't show me new disk devices! How can my cluster do a
probing new disk devices in order to add a new quorum device?

Thanks in advance

Christian

2. Using a modem on NT which is really on Linux

3. Using DiskSuite/Solaris Volume Manager to add new disk to live RAID-0

4. mod_perl API 'user' problem

5. Problems Adding New Disk to RS/6000

6. gtk imagen show HELP

7. Adding a new SCSI Disk with Redhat 8

8. Q: how to restrict symlinks

9. Adding new hard disk to A1000.

10. Adding New Disk on 2.6 (Ultra AXi Motherboard)

11. adding a new disk on system

12. how to add a new disk to FreeBSD ?

13. Adding Newfeeds to a New SCSI Disk