robtest fails and the Drive are not detected and /dev/rmt is empty on the host side

[hamzafayyaz1983]

Hi,

Netbackup version = 6.5
Netbackup server host = solaris 10.

My netbackup server was working fine with the ndmp configuration.

I have reconfigured my netbackup server 6.5  from ndmp to simple host to libraray.

I removed the iscsi cable from the filer to the libarary and
connected my scsi cable from the backupserver to the library.


i have deleted the previus enteries by using the /usr/volmgr...../tpconfig -delete.


Now when i am trying to add my drive and tape library again it throws me an error that ur drive is disabled.

I have followed the steps in the drive configuration guide.

1)  rm -rf /dev/rmt/*
2) st.install
3) reboot -- -r
4) changing the /kernel/drv/st.conf file
5) st.build to regenerate the st.conf file

but my /usr/openv/volmgr/robtest is failing


ROBTEST OUTPUT

Robot Selection
---------------
  1)  TLD 0
  2)  none/quit
Enter choice: 1

Robot selected: TLD(0)   robotic path = /dev/sg/c1t4l1

Invoking robotic test utility:
/usr/openv/volmgr/bin/tldtest -rn 0 -r /dev/sg/c1t4l1

Opening /dev/sg/c1t4l1
user command failed, may be timeout, scsi_pkt.us_reason = 4
user command failed, may be timeout, scsi_pkt.us_reason = 4
mode_sense ioctl() failed: Error 0
----------------------------------------------------------------------------------------


my /dev/rmt/ directory is empty
its not generating any nymber like 4cbn or 7 cbn even after running the scripts st.build and


Long Story Short

Its not recognizing my scsi driver for solaris (done the passthru driver thing).
Its not recognizing my drive





Thanx in advance
Regards,
Hamza

[Will_Restore]

On a Solaris system, sg.build fails to build sg devices for some HBAs.

Introduction:
With certain HBAs and drivers, sg.build will not automatically build /dev/sg devices for fibre channel tapes and robots on Solaris.

Troubleshooting:
By default, sg.build only builds sg devices for HBAs that have "SUNW" listed in the "luxadm -e port" output eg:

# luxadm -e port

Found path to 2 HBA ports

/devices/pci@1e,600000/SUNW,qlc@2/fp@0,0:devctl                    CONNECTED

 

For Resolution, see:
http://seer.entsupport.symantec.com/docs/283852.htm

[Marianne]

Only retry sg.build once /dev/rmt is populated.

[hamzafayyaz1983]

Hi,



I have tried the sg.build once but no results and the /dev/rmt/ is still empty



I have luxadm -e port but no output because i am using the scsi cable not the hba

[Marianne]

sg.build is for /dev/sg only - it will not build /dev/rmt entries. O/S device name is an O/S issue and need to be resolved by your SUN/Solaris engineer/admin.

If you are using scsi, shutdown your server. While at the ok prompt, do 'probe-scsi-all' (you might have to perform a 'reset-all' first and the send a 'break/stop-A' to prevent from booting).
Ensure that the library and tape drive(s) are seen at this level. Only when devices are seen by probe-scsi-all, do a 'boot -r'. Check /dev/rmt.

[hamzafayyaz1983]

Tried it but no lock

[Marianne]

What exactly is not working?
What does probe-scsi-all report?
What is in /var/adm/messages when you boot -r?

[hamzafayyaz1983]

Apr  2 17:12:05 hrsrv01 rmclomv: [ID 211032 kern.error] PSU @ PS0 has FAULTED.
Apr  2 17:27:06 hrsrv01 rmclomv: [ID 211032 kern.error] PSU @ PS0 has FAULTED.
Apr  2 17:42:08 hrsrv01 rmclomv: [ID 211032 kern.error] PSU @ PS0 has FAULTED.
Apr  2 17:57:09 hrsrv01 rmclomv: [ID 211032 kern.error] PSU @ PS0 has FAULTED.
Apr  2 18:12:11 hrsrv01 rmclomv: [ID 211032 kern.error] PSU @ PS0 has FAULTED.
Apr  2 18:27:12 hrsrv01 rmclomv: [ID 211032 kern.error] PSU @ PS0 has FAULTED.
Apr  2 18:42:14 hrsrv01 rmclomv: [ID 211032 kern.error] PSU @ PS0 has FAULTED.
Apr  2 18:57:15 hrsrv01 rmclomv: [ID 211032 kern.error] PSU @ PS0 has FAULTED.

[Will_Restore · Accepted solution]

you seem to have a bad server power supply.

is there nothing in probe-scsi-all or messages regarding tape device?

[hamzafayyaz1983]

I have checked the scsi card from the hardwares guyz and it turned out to be faulty on v240.


Thank You Marianne and others for your help.



The issue is solved.



Regards,
Hamza