Very Computer

Hi

Hopefully this is a faq, but if it is I can't faq'ing find it :PP
Google, you've let me down.

I have a 3510 array on a V440 (Solaris 9 9/05, current Recommended
patch level) and have been setting 7 machines up with Jumpstart. I've
noticed that the disk ID targets appear to be climbing each time I've
jumpstarted the system. Well, perhaps not every time, but I can't work
out why it has some times and not others.

Currently, these are the first disks in my system:

       4. c3t48d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>

       5. c3t49d0 <SUN36G cyl 24620 alt 2 hd 27 sec 107>

Starting at target 48 !

Anything I can do to bring these back to a more sensible number - and
preferably stop them wandering at all ? I don't want to obsess over it,
but I'm writing some system configuration scripts (and associated
documentation) and I'd rather not gamble that things will change should
someone have to do a rebuild in my absence.

Thanks in advance

Andrew

More information:

sccli> show enclosure-status
 Id Chassis Vendor/Product ID    Rev  PLD  WWNN             WWPN
----------------------------------------------------------------------------
 60 0984F2  SUN StorEdge 3510F D 1046 1000 205000C0FF0984F2
215000C0FF0984F2
                                      Topology: loop(a)  Status:
OK

Enclosure Component Status:
     Type Unit Status   FRU P/N   FRU S/N   Add'l Data
------------------------------------------------------------------
      Fan 0    OK       370-6776  GM21NJ     --
      Fan 1    OK       370-6776  GM21NJ     --
      Fan 2    OK       370-6776  GM21NH     --
      Fan 3    OK       370-6776  GM21NH     --
       PS 0    OK       370-6776  GM21NJ     --
       PS 1    OK       370-6776  GM21NH     --
     Temp 0    OK       370-5535  0984F2     temp=30
     Temp 1    OK       370-5535  0984F2     temp=31
     Temp 2    OK       370-5535  0984F2     temp=30
     Temp 3    OK       370-5535  0984F2     temp=31
     Temp 4    OK       370-5535  0984F2     temp=30
     Temp 5    OK       370-5535  0984F2     temp=30
     Temp 6    OK       370-5538  HG109W     temp=35
     Temp 7    OK       370-5538  HG109W     temp=35
     Temp 8    OK       370-5538  HG109W     temp=33
     Temp 9    OK       370-5538  HG109W     temp=33
     Temp 10   OK       370-6776  GM21NJ     temp=33
     Temp 11   OK       370-6776  GM21NH     temp=32
 DiskSlot 0    OK       370-5535  0984F2     addr=48,led=off
 DiskSlot 1    OK       370-5535  0984F2     addr=49,led=off
 DiskSlot 2    OK       370-5535  0984F2     addr=50,led=off
 DiskSlot 3    OK       370-5535  0984F2     addr=51,led=off
 DiskSlot 4    OK       370-5535  0984F2     addr=52,led=off
 DiskSlot 5    OK       370-5535  0984F2     addr=53,led=off
 DiskSlot 6    OK       370-5535  0984F2     addr=54,led=off
 DiskSlot 7    OK       370-5535  0984F2     addr=55,led=off
 DiskSlot 8    OK       370-5535  0984F2     addr=56,led=off
 DiskSlot 9    OK       370-5535  0984F2     addr=57,led=off
 DiskSlot 10   OK       370-5535  0984F2     addr=58,led=off
 DiskSlot 11   OK       370-5535  0984F2     addr=59,led=off

So how come it started at 48 all of a sudden?

Well, I found the answer, and sadly it was that that it didn't start at
48 "all of a sudden". Where I would have bet money it was in the 20's
earlier, I just found the bit in the guide that sets the range of IDs
by the ID switch on the front of the array. The two arrays were set to
ID 2 and 3, and sure enough that started them at 32 and 48.

So how come I didn't spot that sooner? I blame tiredness and stress.

AW (wearing the dunces cap)

  I'm not so sure that's the reason.  Take a look at:

sccli> show channels
Ch  Type    Media   Speed   Width  PID / SID
--------------------------------------------
 0  Host    FC(L)   2G      Serial  40 / 41
 1  Host    FC(L)   2G      Serial  43 / 42
 2  Drive   FC(L)   2G      Serial  14 / 15
 3  Drive   FC(L)   2G      Serial  14 / 15
 4  Host    FC(L)   2G      Serial  44 / 45
 5  Host    FC(L)   2G      Serial  47 / 46
 6  Host    LAN     N/A     Serial  N/A / N/A

  I'm pretty sure the numbers you're seeing are the channel definitions
above.  My example does not have channels 32 and 48 defined but
your's almost certainly does.  The ID switch on the front of the arrays
is for connecting JBOD disk trays to a controller (or dual controller)
array.  

- Charlie

I'm currently scratching my head over this too.
I have been monitoring a couple of systems with several 3510-units for some
time now and I find that these numbers aren't even constant for the same
system. I'm not sure exactly when the IDs change because this hasn't been
24/7 monitoring. I'm sitting with an example before me whith 2 large
HA-clusters and four 3510 units. Every single ID on all units has changed
since I probed the systems in August.

Quote:> Well, I found the answer, and sadly it was that that it didn't start at
> 48 "all of a sudden". Where I would have bet money it was in the 20's
> earlier, I just found the bit in the guide that sets the range of IDs
> by the ID switch on the front of the array. The two arrays were set to
> ID 2 and 3, and sure enough that started them at 32 and 48.

> So how come I didn't spot that sooner? I blame tiredness and stress.

> AW (wearing the dunces cap)