Raid Manager cannot detect new A1000

[yuroy]

Hi All,
I'm trying to configure an A1000 disk array to an E420R (Solaris 8) machine. The A1000 has been configured before on an older machine, but when I plugged it on the E420R, Raid Manager cannot see the A1000. I ran "format" and the disk were displayed. I selected the disk and this error was displayed:
-----------------------------------------
Specify disk (enter its number): 5
selecting c3t5d1
[disk unformatted]
Disk not labeled. Label it now? y
Warning: error writing VTOC.
Illegal request during read
ASC: 0x91 ASCQ: 0x50
Warning: error reading backup label.
Illegal request during read
ASC: 0x91 ASCQ: 0x50
Warning: error reading backup label.
Illegal request during read
ASC: 0x91 ASCQ: 0x50
Warning: error reading backup label.
Illegal request during read
ASC: 0x91 ASCQ: 0x50
Warning: error reading backup label.
Illegal request during read
ASC: 0x91 ASCQ: 0x50
Warning: error reading backup label.
Warning: no backup labels
Write label failed
----------------------------------------
Would anyone know how to resolve this?

thanks a lot.
Roy

 

[KenCunningham]

Ah, the A1000, my favourite (not!). Have you patched the box so that Raid Manager is as up-to-date as possible? Can you run the raid manager commands from the command line - if not, do you get any error messages?

 

[yuroy]

Hi,
Yes, I can run Raid Manager on command line. I'm on ver 6.22.01.
I already have an existing A1000 attached, just wanted to add a new one.

I noticed that when I ran 'format' on command line, the path (or driver?) for the working A1000 is different. Pls. see below: 2 & 3 are the working A1000 and 4 & 5 is the new A1000 that I'm having problem with. If you notice for 2 & 3, the path is "/pseudo/rdnexus@1/rdriver@x,x" while for 4 & 5, path is "/pci@1f,4000/scsi@4/sd@x,x" just like for internal disks 0 & 1.

Any chance this could be the reason? and if so, how do I modify this?

Thanks in advance.
Roy
------------------------------------
0. c0t0d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@1f,4000/scsi@3/sd@0,0
1. c0t1d0 <SUN9.0G cyl 4924 alt 2 hd 27 sec 133>
/pci@1f,4000/scsi@3/sd@1,0
2. c1t5d0 <Symbios-StorEDGEA1000-0205 cyl 43084 alt 2 hd 64 sec 64>
/pseudo/rdnexus@1/rdriver@5,0
3. c1t5d1 <Symbios-StorEDGEA1000-0205 cyl 43084 alt 2 hd 64 sec 64>
/pseudo/rdnexus@1/rdriver@5,1
4. c3t5d0 <Symbios-StorEDGEA1000-0301 cyl 43084 alt 2 hd 64 sec 64>
/pci@1f,4000/scsi@4/sd@5,0
5. c3t5d1 <Symbios-StorEDGEA1000-0301 cyl 43084 alt 2 hd 64 sec 64>
/pci@1f,4000/scsi@4/sd@5,1
---------------------------------------

 

[KenCunningham]

Yes, I think that's your problem - my three A1000's are all pseudo entries like your 'good' ones above. As a matter of interest, are the SCSI interfaces for both A1000's the same? If they are (and you are able to do so), I think I'd try to remove the spurious entries and do a reconfigure boot to try to get the OS to recognise the device correctly.

 

[yuroy]

Hi,
Thanks for the reply.
How do I remove those entries?
I've tried rebooting with "boot -r" but I still get the same entries.
Maybe I do need to remove those entries and reboot it again. Pls. advise how I can remove those entries?

Thanks again.
Roy

 

[KenCunningham]

For further info, the cards we are using to connect the A1000's are X6541 - PCI dual channel differential ultra scsi host adapter. A bit of Googling confirms that the mistake often encountered with A1000's involves using a non-diff device for the interface.

 

[KenCunningham]

Sorry - posting at the same time. You could try disconnecting the A1000 and doing another boot -r whilst it is not connected, then conect it up again and do another boot -r. A bit long-winded but it could be effective. If that doesn't work, others may be able to help with removing the entries manually - I've never had to do it, fortunately.

 

[yuroy]

Thanks. I'll try doing that once I get people off the server. I might have to do that late tonight or early tomorrow morning. I'll update you on the result.

thanks again.

 

[KenCunningham]

Good luck - have you confirmed that you SCSI card is appropriate?

 

[KenCunningham]

Doing a little more digging reveals that devfsadm might do what you require without a reboot. There is a -C option which cleans up dangling devices, apparently. I've never used it, but take a look at the man page before doing anything drastic!

 

[yuroy]

Hi,
It's working now. Somehow I had to use the /reconfigure trick to have it recognize /pseudo. "boot -r" wasn't able to find the /pseudo. Raid Manager now recognizes it and am now able to configure it.

Many thanks for the help.
Roy

 

[KenCunningham]

Roy, excellent news. Thinking about it, I have heard before that boot -r doesn't always pick up every config change, wish I'd remembered that yesterday. Anyway, all's well that ends well! Cheers.