Sun Storage A3500 SCSI Array
 
Sun Storage A3500 FC Array
 

GCS>Sun Microsystems>Storage - Disk>Modular Disk - Other
 

  ------------------------------------------------------------------------
* probe-scsi-all *
The SRC/P card must be visible at the OK prompt in order
for us to use it. It reports quite differently than its
counterpart, the SYMBIOS controller used with rm6. Below
are some examples of what you should see (and not see).
If the card is successfully installed and functioning, it
will be seen by probe-scsi-all. In the following example, a
single raid card has been installed in slot 1 of an e450.
There are no disk drives installed behind the card.
probe-scsi-all should report 3 paths (one for each channel)
to the card. Notice that mscsi is in the device path. This
is a unique identifier for the card.
/pci@6,4000/pci@3/scsis@4/mscsi@2,0
/pci@6,4000/pci@3/scsis@4/mscsi@1,0
/pci@6,4000/pci@3/scsis@4/mscsi@0,0
* show-devs *
Similarly, the "show-devs" command will yield an identical
result.
/pci@6,4000/pci@3/scsis@4/mscsi@2,0
/pci@6,4000/pci@3/scsis@4/mscsi@1,0
/pci@6,4000/pci@3/scsis@4/mscsi@0,0
/pci@6,4000/pci@3/scsis@4/mscsi@2,0/tape
/pci@6,4000/pci@3/scsis@4/mscsi@2,0/disk
/pci@6,4000/pci@3/scsis@4/mscsi@1,0/tape
/pci@6,4000/pci@3/scsis@4/mscsi@1,0/disk
/pci@6,4000/pci@3/scsis@4/mscsi@0,0/tape
/pci@6,4000/pci@3/scsis@4/mscsi@0,0/disk
Now, let's add some disk drives behind the card. The drives
will show up with probe-scsi-all so long as they are NOT
part of a Logical Storage Unit. As soon as they are
incorporated into an LSU they will no longer show up in a
probe. The following probe-scsi-all identifies two drives
installed behind the raid card.
/pci@6,4000/pci@3/scsis@4/mscsi@1,0
Target 0
Unit 0 Disk Fujitsu MAB3091S SUN9.0G1705
Target 1
Unit 0 Disk Fujitsu MAB3091S SUN9.0G1705
After creating some Logical Storage Units the output at the
OK prompt dramatically changes. probe-scsi-all will NO
longer report any disk drive which has been incorporated
into an LSU. Instead, probe will now identify ONLY the LSU.
In the following example, a RAID5 device was created from 5
drives behind the card. Only the LSU is reported now. The
drives which make up the LSU are gone!
/pci@6,4000/pci@3/scsis@4/mscsi@1,0
Target 0
Unit 0 Disk DPT RAID5 210C
* obdiag *
There are two more diagnostics available at the OK prompt
to verify the card is good. One is obdiag which will
	display a menu of all the devices seen by the PROM. They
	appear in a numerical order.
An example is provided below.
OK obdiag
1. SUNW,envctrl@14,60000
2. SUNW,ffb@1d,0
3. ebus@1,1
4. eccp@14,3043bc
5. eeprom@14,0
6. more devices....
10. scsis@4
11. se@14,400000
12. su@14,3062f8
13. su@14,3083f8
Notice that number 10 is scsis@4. From the previous
probe-scsi-all, we can see that scsis@4 is in the path to
the hardware raid card. It is this "number" that we can use
to test the device.
command> test 10
Hit the spacebar to interrupt testing
Testing /pci@6,4000/pci@3/scsis@4
................................. passed
Hit any key to return to the main menu
* show-pci-config *
The last diagnostic we have is show-pci-config. We can
actually see if the card is "busted" with this diagnostic.
Broken cards will put themselves into a "blinked" state.
This state is identifiable with show-pci-config.
OK show-pci-config /pci@6,4000/pci@3/scsis@4
PCI Header for Bus #xx, Dev #x, Func # x
Vendor ID = 1044
Device ID = a501
Command Reg = 0
Status Reg = 280
Revision ID = 3
Class Code = ebc14
Cache Line Size = 0
Latency Timer = 0
Header Type = 1
etc.......
OK
It's the "class code" which lets us know if the card is
good. In the above example, we see a class code of "ebc14".
If the card was healthy, we would have seen a class code of
"e0000". The "e" will always be present, the "bc" means the
card is "blinked" or failed, and the last two digits, (14),
are the blink failure code.

This solution has no attachment