SUMMARY: Devices either missing or incorrect after boot -r

From: Andrew Harvey <>
Date: Fri Feb 22 2002 - 17:37:18 EST
System: Ultra-4 E450 Solaris 2.8 JNI FCI-1063 (Tachyon) NetBackup 3.4.1

    Last week I updated the JNI driver from the 2.7 to the FCI-1063 driver with support for 2.8, and since then I have not been able
to see any of my drives using sgscan, and mt -f /dev/rmt/2cbn stat produces (No such device or address) 
Feb 19 19:24:28 ns0 avrd[277]: Fatal open error on DRIVE1 (device 0, /devices/pci@4,4000/fibre-channel@2/sg@5,0:raw), errno = 19 (No such device), DOWN'ing it

Actions I took after speaking to to you.

I'm at home and writing this from memory so there may be a few discrepancies but as I 
recall, this is now the situation.

[1] cd /dev/rmt && rm *
[2] rm /devices/pci@4,4000/fibre-channel@2/sg@7,0:raw
[3] vi /kernel/drv/sd.conf and comment out 3 lines similar to
name="sd" class="scsi" target=5 lun=0 wwpn="1234567891011A2";
name="sd" class="scsi" target=6 lun=0 wwpn="123456789101187";
name="sd" class="scsi" target=7 lun=0 wwpn="123456789101164"; 
[4] rm /kernel/drv/sg.conf
[5] rm /kernel/drv/st.conf
[6] rm_dev sg
[7] sg_install
[8] all -mt 15 -ml 1
[9] remove all but the dtf line and the dtf-data & hex number lines in st.conf
[10] add the new 0-6 8-15 st.conf entries to /kernel/drv/sd.conf
       (this file is now small and neat!)
[11] replace the sg links with the 0-6 8-15 in /etc/
[12] devfsadm && drvconfig && tapes 

        (i) sgscan now shows all 3 drives and the robot
        (ii) I think I remember /dev/rmt/2 is now simlinked to
        Anyhow, when I left the office around 8.30, all drives were accessible 
       there were no SCSI errors and one of my backup jobs was completed in 2.5 mins


Andrew Harvey
Help Desk
sunmanagers mailing list
Received on Mon Feb 25 02:29:05 2002

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:42:35 EST