North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: Sparc Storage Array

  • From: Jake Khuon
  • Date: Mon Apr 03 07:44:23 2000
  • Action:
  • Dcc:
  • Expires:

### On Sun, 2 Apr 2000 23:01:36 -0400, Jim Mercer <[email protected]>
### casually decided to expound upon [email protected] the following thoughts
### about "Sparc Storage Array":

JM> quick attempt to draw on the accumulated knowledge herein.
JM> 
JM> i'm looking to gain som help in getting a fiber channel attached
JM> sparc storage array (100?) working with a sparc 20, solaris 7.
JM> 
JM> if you have worked with this, contact me directly.

I have a Sun SSA model 100/1010 (least I believe that's what it says on the
back).  What kind of fibrechannel board do you have installed on the SS20
and the SSA?  I believe the FC interfaces on the SSA are only available as
the older non-arbitrated kind.  You may use the older FC Sbus board or the
newer FC/FC-AL (it will revert to standard FC) Sbus board.

I have the Sun FC/FC-AL which shows up at soc0 ("Sun Serial Optical
Controller" FCS Host Adapter) and the drive array is dual attached and shows
up under pln0 and pln1.  Solaris7 should have all the drivers installed by
default although you may need to upgrade the firmware on the SSA.  I get the
following warning:

Mar 22 12:29:48 Espresso.NEEBU.Net unix: ID[SUNWssa.soc.link.1010] soc0: port 0 WWN 740c8d: message:  SSA100 V3.6 (031896) Mon Mar 18 19:57:51 1996  
Mar 22 12:29:49 Espresso.NEEBU.Net unix: ID[SUNWssa.soc.link.1010] soc0: port 1 WWN 740c8d: message:  SSA100 V3.6 (031896) Mon Mar 18 19:57:51 1996  
Mar 22 12:29:49 Espresso.NEEBU.Net unix: NOTICE: pln0: Old SSA firmware has been detected (Ver:3.6) : Expected (Ver:3.12) - Please upgrade
Mar 22 12:29:49 Espresso.NEEBU.Net unix: NOTICE: pln1: Old SSA firmware has been detected (Ver:3.6) : Expected (Ver:3.12) - Please upgrade

However, [1] I don't exactly know how to upgrade the firmware... I'm
assuming it can be done via some kind of flash-routine and [2] I'm able to
use the array fine although the small non-problem I describe below may be
attributable to old firmware... I dunno.

Here's an example output from sysinfo on my system (Ultra2 running Sol7):

SUNW,soc0 is a "Sun Serial Optical Controller" FCS Host Adapter
  SUNW,pln0 is a "Sun SPARCstorage Array" Host Bus Adapter driver
    c1t0d0 (ssd0) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive
    c1t0d1 (ssd1) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t0d2 (ssd2) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t0d3 (ssd3) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t0d4 (ssd4) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive
    c1t1d0 (ssd5) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t1d1 (ssd6) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t1d2 (ssd7) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t1d3 (ssd8) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive
    c1t1d4 (ssd9) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t2d0 (ssd10) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t2d1 (ssd11) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t2d2 (ssd12) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t2d3 (ssd13) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive
    c1t2d4 (ssd14) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t3d2 (ssd17) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t3d4 (ssd19) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t4d2 (ssd22) is a "CONNER CFP1060E SUN1.05" SCSI 1.2 GB disk drive
    c1t4d4 (ssd24) is a "IBM DCAS32160SUN2.1G" SCSI 2.0 GB disk drive
    c1t5d2 (ssd27) is a "FUJITSU MAB3045S SUN4.2G" SCSI 4.0 GB disk drive
    c1t5d4 (ssd29) is a "FUJITSU MAB3045S SUN4.2G" SCSI 4.0 GB disk drive
  SUNW,pln1 is a "Sun SPARCstorage Array" Host Bus Adapter driver
    c2t0d0 (ssd30) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive
    c2t0d4 (ssd34) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive
    c2t1d3 (ssd38) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive
    c2t2d3 (ssd43) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive
    c2t3d3 (ssd48) is a "CONNER CFP1080E SUN1.05" SCSI 1.3 GB disk drive

Note that for some reason, the OS didn't configure all the drives under pln1
as it did under pln0.  The pln0 interface should expose ssd0 through ssd29
and the pln1 interface should expose ssd30 through ssd59.  I probably need
to muck with /kernel/drv/ssd.conf or something to get all the drives to show
visibility off both the interfaces simultaneously.  However, if you don't
run Veritas LVM with alternate/multi device-pathing then this shouldn't be
of concern.  I'm currently only using Solstice DiskSuite which doesn't
support multipathing so I just ignore pln1 devices altogether.  A better
idea is to probably just singly-attach with only one interface so the OS
won't get confused.  Also, DiskSuite kind of goes a little wonky when you
dual-attach and sees one interface attached to a FC controller but sees the
other one as attached to a standard SCSI controller.  I just ignored all the
disks that DiskSuite claims is attached to the faux-SCSI controller.  This
might be a firmware issue.

Additional note:  I had this same setup running on a SS5 so I don't think
there's much of a difference issue between sun4m and sun4u with respect to
the SSA100.

Hope this helps...


--
/*====================[ Jake Khuon <[email protected]> ]======================+
 | Network Systems Engineer, Net-Eng/NSM      /~_ |_ () |3 /-\ |_ _ GLOBAL |
 | VOX: +1(408)543-4828  Fax: +1(408)543-0074 \_| C R O S S I N G   CENTER |
 +===============[ 141 Caspian Court, Sunnyvale, CA  94089 ]===============*/