Very Weird SCSI Id Problem

From: Greg Wilson (gwilson_at_connect.com.au)
Date: 02/02/05

  • Next message: John DiMarco: "Sun Managers Frequently Asked Questions (FAQ)" 
    Date: Wed, 2 Feb 2005 16:21:50 +1100
To: sunmanagers@sunmanagers.org

    okies

    Hardware:

    1 x Sun 420R
    2 x A1000
    2 x Diff SCSI cards

    Senario:

    All working fine with 420R, 1x A1000 and 1x Diff SCSI

    We need to add another A1000 as it was saturating the SCSI BUS

    So 8 months ago we added another Diff SCSI card and A1000 and boot -r

    The new card appeared as

           3. c69t5d0 <Symbios-StorEDGEA1000-0301 cyl 52473 alt 2 hd 128 sec 64>
              /pci@1f,4000/scsi@2/sd@5,0

    mmm ive added card before and the only time i have seen c69t5* is when we
    used a 5200 on an enterprise machine.

    I though no dig deal of it and just made another raid volume and everything
    returned to normal.

    Then yesterday we lost the battery of the new A1000. This is where the weird
    stuff began

    we inserted the new battery and rebooted.

    when i ran the command /etc/raid/bin/raidutil -c c69t5d0 -R to reset the
    battery age it said Device not found: c69t5d0

    hmm.. i ran the command format

    <snip>
      titan[/etc/raid/bin]# format
      Searching for disks...done

      AVAILABLE DISK SELECTIONS:
           0. c0t0d0 <SUN18G cyl 7506 alt 2 hd 19 sec 248>
              /pci@1f,4000/scsi@3/sd@0,0
           1. c0t1d0 <SUN18G cyl 7506 alt 2 hd 19 sec 248>
              /pci@1f,4000/scsi@3/sd@1,0
           2. c1t5d0 <Symbios-StorEDGEA1000-0301 cyl 17232 alt 2 hd 64 sec 64>
              /pseudo/rdnexus@1/rdriver@5,0
           3. c69t5d0 <Symbios-StorEDGEA1000-0301 cyl 52473 alt 2 hd 128 sec 64>
              /pci@1f,4000/scsi@2/sd@5,0
      Specify disk (enter its number): ^D
    <snip>

    then i ran the command /etc/raid/bin/lad

    <snip>
      titan[/etc/raid/bin]# lad
      c1t5d0 1T00844921 LUNS: 0
      c2t5d0 1T10514421 LUNS: 0
    <snip>

    hmm..

    So the problem is Raidmanager thinks the SCSI Id is c2t5 but the OS sees
    its as c69t5*

    has anyone else seen this. im kinda worried about this now

    <snip>

      titan[/etc/raid/bin]# raidutil -c c69t5d0 -i
      Device not found: c69t5d0

      raidutil failed!
      titan[/etc/raid/bin]# format
      Searching for disks...done

      AVAILABLE DISK SELECTIONS:
           0. c0t0d0 <SUN18G cyl 7506 alt 2 hd 19 sec 248>
              /pci@1f,4000/scsi@3/sd@0,0
           1. c0t1d0 <SUN18G cyl 7506 alt 2 hd 19 sec 248>
              /pci@1f,4000/scsi@3/sd@1,0
           2. c1t5d0 <Symbios-StorEDGEA1000-0301 cyl 17232 alt 2 hd 64 sec 64>
              /pseudo/rdnexus@1/rdriver@5,0
           3. c69t5d0 <Symbios-StorEDGEA1000-0301 cyl 52473 alt 2 hd 128 sec 64>
              /pci@1f,4000/scsi@2/sd@5,0
      Specify disk (enter its number):

      titan[/etc/raid/bin]# lad
      c1t5d0 1T00844921 LUNS: 0
      c2t5d0 1T10514421 LUNS: 0

      titan[/etc/raid/bin]# raidutil -c c2t5d0 -i
      LUNs found on c2t5d0.
      LUN 0 RAID 5 209901 MB

      Vendor ID Symbios
      ProductID StorEDGE A1000
      Product Revision 0301
      Boot Level 03.01.04.00
      Boot Level Date 04/05/01
      Firmware Level 03.01.04.81
      Firmware Date 10/17/03
      raidutil succeeded!

      titan[/etc/raid/bin]# df -kFufs
      /dev/md/dsk/d120 211663336 76905077 132641626 37% /export/01
      
      titan[/etc/raid/bin]# metastat d120
      d120: Mirror
        Submirror 0: d121
          State: Okay
        Pass: 1
        Read option: roundrobin (default)
        Write option: parallel (default)
        Size: 429850624 blocks

        d121: Submirror of d120
        State: Okay
        Size: 429850624 blocks
        Stripe 0:
            Device Start Block Dbase State Hot Spare
            c69t5d0s2 8192 Yes Okay

      titan[/etc/raid/bin]# metadb -i
            flags first blk block count
         a m p luo 16 1034 /dev/dsk/c0t0d0s7
         a p luo 1050 1034 /dev/dsk/c0t0d0s7
         a p luo 2084 1034 /dev/dsk/c0t0d0s7
         a p luo 16 1034 /dev/dsk/c0t1d0s7
         a p luo 1050 1034 /dev/dsk/c0t1d0s7
         a p luo 2084 1034 /dev/dsk/c0t1d0s7
         a p luo 16 1034 /dev/dsk/c69t5d0s2
     
    <snip> 

    -- 
Greg Wilson	Senior System Administrator	greg.wilson@aapt.com.au
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
  • Next message: John DiMarco: "Sun Managers Frequently Asked Questions (FAQ)"

    Relevant Pages

    • Re: Linux 2.6.4 - Card Reader not working
      ... > it could see all LUNs on the device. ... But when I plug in a card, no "part1" appears on the appropriate LUN. ... I now needed to make usb-storage as well as the scsi stuff as modules in ... I attached the dmesg output from the modprobing of the usb ...
      (Linux-Kernel)
    • Re: Unipod woes
      ... Expansion card 0: Castle Technology Storm DMA32 SCSI interface ...
      (comp.sys.acorn.hardware)
    • Re: FS: Lots of computer parts - Redhill
      ... - eg add 75p per card for postage. ... two small SCSI drives and a SCSI controller card, ... Either add memory for basic PC or simply strip out ... Either way a snip at £10 ...
      (uk.local.surrey)
    • Re: Ultra 10 - no output to monitor
      ... DIMM sockets -- bank 0 being the pair closest to the edge of the board, ... which is *not* a PCI bus card.) ... pull them, and the Creator-3D framebuffer, and perhaps half of the DIMMs ... The SCSI card was ...
      (comp.sys.sun.hardware)
    • Re: CF7 and P-box y-cable tests in progress...
      ... If I print to PIO/2, the RS232 card lights up, but ... SCSI issues were pretty bad during this testing - consistent lockups, ... Checked SCSI connections (they ... CF7 operations seemed unaffected. ...
      (comp.sys.ti)