SUMMARY: V880 WWN problem

From: Joohyun Cha <zoo11_at_hst.co.kr>
Date: Thu Jul 08 2004 - 00:55:56 EDT
Hi. Managers.

I'm sorry for late summary. Thanks for your reply to
Jon Hudson
Marian Dragomir
Hirchert, Dale
Bill R. William
Westerdale, John
Pascal Grostabussiat

Please forgive me if I forgot someone.

My Original Question and responses are below.

Conclusion: It's pretty normal when disks are dual ported. BUT mine has not
dual ported disks. I suspect that first configuration of this system has
dual ported disks but for some reason one port has been disabled or removed
physically. I think If this happened it could be. But I'm pretty not sure.

Some people stated that this is quite often on some Fujitsu or Seagate disk.

And also I want to ask some papers or references on the net about this topic.

Best regards,


------------------------------------------------------------------------
Hi. Gurus

I have V880 with Solaris 8 02/04(117350-02) release.
While I was playing with it I felt that something's wrong. That is
each disk's WWN outputed by luxadm isn't same with output of 'ls /dev/dsk
or format'
Actually WWN of 'ls /dev/dsk' is exactly +1 of luxadm's. (please see below)
but system is ok so far. What's the problem? is it normal? if not normal
what action should I take?

Please give me some advice. (it's a production system)

# luxadm probe
Found Enclosure:
SUNWGS INT FCBPL   Name:FCloop   Node WWN:50800200001d3500   Logical
+Path:/dev/es/ses0

# luxadm display 50800200001d3500

                SUNWGS INT FCBPL
                 DISK STATUS

SLOT   DISKS             (Node WWN)
0      On (O.K.)         500000e01063f350
1      On (O.K.)         500000e01063f340
2      On (O.K.)         500000e010634c30
3      On (O.K.)         500000e010640c40
4      On (O.K.)         500000e010640d00
5      On (O.K.)         500000e010640c60
6      On (Login failed)
...
...

# format
AVAILABLE DISK SELECTIONS:

       0. c1t0d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
          /pci@8,600000/SUNW,qlc@2/fp@0,0/ssd@w500000e01063f351,0
       1. c1t1d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
          /pci@8,600000/SUNW,qlc@2/fp@0,0/ssd@w500000e01063f341,0
       2. c1t2d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
          /pci@8,600000/SUNW,qlc@2/fp@0,0/ssd@w500000e010634c31,0
       3. c1t3d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
          /pci@8,600000/SUNW,qlc@2/fp@0,0/ssd@w500000e010640c41,0
       4. c1t4d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
          /pci@8,600000/SUNW,qlc@2/fp@0,0/ssd@w500000e010640d01,0
       5. c1t5d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
          /pci@8,600000/SUNW,qlc@2/fp@0,0/ssd@w500000e010640c61,0

Specify disk (enter its number):

------

are they dual ported disks? They may have given each path a separate WWN. Like
+for dual port cars the WWN will start with a 20 but then each port will start
+with 10.

I'm taking a guess here as I don't have access to my v880 here, but I can check
+tomorrow.
No way for me to put a analyzer on the FCAL loop as it's internal. But I have
+you checked the disks? They are labled.

------
Hello!
On some Fujitsu or SEAGATE disks, this is a normal situation.
This disk drives use different WWN scheme than traditionally seen.
If you have dual path to disks and you have VxVM, you must add Device
Discovery Layer support for them.

Example:

Find Disk Vendor ID:

#/etc/vx/diag.d/vxdmpinq /dev/rdsk/c0t1d0s2

Here you will see VID (Vendor Id) and Product id (PID).
Now you must run the following commands:

#vxddladm addjbod vid=... pid=....
#vxdctl enable

If the device is FUJITSU , you must include the length too.
#vxddladm addjbod vid=FUJITSU pid=... length=10

Best regards,


--------

My experience with WWN numbers are that they generally are annoying
messages. I've only seen this when disks have been swapped around from
one server to another, but I've never seen any performance problems or
issues associated with the annoying message. You don't mention getting
an "Invalid WWN" message during bootup, are you getting any error
messages?

The only way I've been able to "reset" the WWN number is by reloading
the box. Perhaps others have givne you a more logical answer than what I
have said here.

-------
I am not an expert, but lately I have been delving into this very
thing.  Here's some hints...

As you have discovered 'luxadm probe'  will show your enclosure.
On my v880 I have TWO Logical Path: entries -- ses0 and ses1.

In addition to the enclosure's WWN, you can use 'luxadm display
/dev/es/ses0' (or ses1) to display the enclosure details.

Here's the part which will somewhat clear things up:
For each of the "On (O.K.)" disks listed under:
  SLOT   DISKS             (Node WWN)
  0      On (O.K.)         500000e01063f350
  ...
do a 'luxadm display' fo the WWN for the "Node WWN" entry.  For the
disk in slot 0 you would do: luxadm display 500000e01063f350
This will give you all sorts of details about that disk including the
PATHS to that disk.

The disk drives in (my) v880 are DUAL PORTED Fibre Connect drives!
This means that there are TWO pathways into the device.
The last lines of my 'luxadm display' look something like this:
  Path(s):
  /dev/rdsk/c1t0d0s2
  /devices/pci@8,600000/SUNW,qlc@2/fp@0,0/ssd@wXXXXXXXXXXXXXXXXX,0:c,raw
  /dev/rdsk/c2t0d0s2
  /devices/pci@9,600000/pci@1/SUNW,qlc@4/fp@0,0/ssd@wXXXXXXXXXXXXXXXXX,0:c,raw
where the 'XXXXX...' are the WWN for each DEVICE.
In this case, the output means that /dev/rdsk/c1t0d0s2 and
/dev/rdsk/c2t0d0s2 are addresses for the same physical device.

The thing to keep in mind is that each of these things is slightly
different:
*       WWN of the disk's "(Node WWN)" entry from display ses.
*       WWN each of the paths (/devices/...) entries for the disk.
For example I have a disk
        Node WWW:       500000e010366d20
        WWN(Node):      500000e010366d20
        WWN(Port B):    500000e010366d21
with devices:
  Path(s):
  /dev/rdsk/c1t0d0s2
  /devices/pci@8,600000/SUNW,qlc@2/fp@0,0/ssd@w500000e010366d21,0:c,raw
  /dev/rdsk/c2t0d0s2
  /devices/pci@9,600000/pci@1/SUNW,qlc@4/fp@0,0/ssd@w500000e010366d22,0:c,raw
Observe that for this one physical disk device I have THREE different
WWN assignments -- two of them for the paths whereby I address the
disks.

Biggie hint:  On my v880, in most cases the WWN numbers for the
/devices for each disk address differ in  the last digit:
  500000e010366d21
  500000e010366d22
however, this is not always true!  We had a disk replaced and the new
drive came online with the difference in the FIRST TWO digits!
It has 21000004c... and 22000004c... with the rest of the WWN the
same.  It doesn't really matter where they vary as long as each in
unique within Ghe system.

One final thing to remember:  As I already mentioned in the example
/dev/rdsk/c1t0d0s2 and /dev/rdsk/c2t0d0s2 are addresses for the same
physical device.  This means that anything I do to /dev/rdsk/c1t0d0s2
I am also doing to /dev/rdsk/c2t0d0s2 and vice versa.

Hope this helps.

---------

Hello Joohyun,

The WWN for Fujitsu disks are strange as you note.

The disk is labeled with WWN ending in zero.

If the disk has dual paths, the first path is named with the last character
of 1, the second with last char of 2.

I filed this as a bug a year ago, and Sun's response was- "Its not a bug".

Would rather see the vfstab entries point to a simple device that has a
major / minor number, but thats a topic for another day :^)

Go figure.

Best,


-------


There is no problem, I think you are just mixing WWN PORT and WWN NODE.

-- 
------------------------------------------------------------------------------
From:
main(){int a=122,j=11;while(a>-50){a=a>0?a:111;printf("%c",a);a=j==49?46:a-j;
j*=j;j=j%2?j:7;j*=j%8?1:0;}printf("%c1\n",49);}
------------------------------------------------------------------------------
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Received on Thu Jul 8 00:56:45 2004

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:43:34 EST