ioscan shows 27 tape drives when we only have 2

[Martin Corner]

Hi,

We have an rp4440, EVA3000 and an MSL 6030 all connected via 2 Cisco MDS 9216 fibre switches. Our Cisco expert assures me that all zones are set-up correctly We have secure path installed but an ioscan -fnC tape shows 27 tape drives when we do only have 2. The ioscan does show the correct number of LUN's that are presented to the unix box from the SAN.

Even though it shows 27 none of them work.

I have included a copy of the ioconfig, tdlist and spmgr display

Thanks in Advance

[D Block 2]

Sue,

have you looked at running:
insf -C tape -e

[Martin Corner]

Hi,

Tried the insf command and got the following message:

insf: Unable to load the driver: drmfglinsf: Unable to load the driver: gvid_him
_fglrxinsf: Unable to load the driver: drmfglrxinsf: Installing special files fo
r stape instance 0 address 0/2/1/0.69.0.0.0.0.2
insf: Installing special files for stape instance 17 address 0/2/1/0.69.0.0.0.0.
3
insf: Installing special files for stape instance 1 address 0/2/1/0.69.0.0.0.1.2

insf: Installing special files for stape instance 18 address 0/2/1/0.69.0.0.0.1.
3
insf: Installing special files for stape instance 2 address 0/2/1/0.69.0.0.0.2.2

insf: Installing special files for stape instance 19 address 0/2/1/0.69.0.0.0.2.
3
insf: Installing special files for stape instance 3 address 0/2/1/0.69.0.0.0.3.2

insf: Installing special files for stape instance 20 address 0/2/1/0.69.0.0.0.3.
3
insf: Installing special files for stape instance 4 address 0/2/1/0.69.0.0.0.4.2

insf: Installing special files for stape instance 21 address 0/2/1/0.69.0.0.0.4.
3
insf: Installing special files for stape instance 5 address 0/2/1/0.69.0.0.0.5.2

insf: Installing special files for stape instance 22 address 0/2/1/0.69.0.0.0.5.
3
insf: Installing special files for stape instance 6 address 0/2/1/0.69.0.0.0.6.2

insf: Installing special files for stape instance 23 address 0/2/1/0.69.0.0.0.6.
3
insf: Installing special files for stape instance 7 address 0/2/1/0.69.0.0.0.7.2

insf: Installing special files for stape instance 24 address 0/2/1/0.69.0.0.0.7.
3
insf: Installing special files for stape instance 8 address 0/2/1/0.69.0.0.0.8.2

insf: Installing special files for stape instance 25 address 0/2/1/0.69.0.0.0.8.
3
insf: Installing special files for stape instance 9 address 0/2/1/0.69.0.0.0.9.2

insf: Installing special files for stape instance 26 address 0/2/1/0.69.0.0.0.9.
3
insf: Installing special files for stape instance 10 address 0/2/1/0.69.0.0.0.10
.2
insf: Installing special files for stape instance 27 address 0/2/1/0.69.0.0.0.10
.3
insf: Installing special files for stape instance 11 address 0/2/1/0.69.0.0.0.11
.2
insf: Installing special files for stape instance 28 address 0/2/1/0.69.0.0.0.11
.3
insf: Installing special files for stape instance 12 address 0/2/1/0.69.0.0.0.12
.2
insf: Installing special files for stape instance 29 address 0/2/1/0.69.0.0.0.12
.3
insf: Installing special files for stape instance 13 address 0/2/1/0.69.0.0.0.13
.2
insf: Installing special files for stape instance 30 address 0/2/1/0.69.0.0.0.13
.3
insf: Installing special files for stape instance 14 address 0/2/1/0.69.0.0.0.14
.2
insf: Installing special files for stape instance 31 address 0/2/1/0.69.0.0.0.14
.3
insf: Installing special files for stape instance 15 address 0/2/1/0.69.0.0.0.15
.2
insf: Installing special files for stape instance 32 address 0/2/1/0.69.0.0.0.15
.3
insf: Installing special files for stape instance 16 address 0/2/1/0.69.0.255.0.
0.2
insf: Installing special files for stape instance 33 address 0/2/1/0.69.0.255.0.
0.3
insf: Unable to unload the driver: drmfglinsf: Unable to unload the driver: gvid
_him_fglrxinsf: Unable to unload the driver: drmfglrx

[Ashwani Kashyap]

Only two tape LUNs are shown but the target keeps changing , producing multiple devices .

I would take one more detailed look at the zoning . My be your card /dev/td0 is in multiple zones .

I would also delete all refenrces to the tape drives by running rnsf aginst each of them and if possible reboot and let the system do an insf -e while booting and see if you get the correct devices .

However the error of unable to load/unload drivers during insf -e has to be taken care of .what type of drivers are they . Do you have the correct drivers and patches for your FC cards .

I would also run an fcmsutil agains /dev/td0
#fcmsutil /dev/td0 get remote all
and see if you can query the targets with the correct WWN and their status is ready .

[Steven E. Protter]

You may need to clean out the old device files and start fresh.

cd /etc
mv ioconfig ioconfig.sav

boot the box.

You'll be prompted at boot time to rebuild the drivers.

ioinit I believe.

You'll need to insf -C afterwards.

This is a procedure that can make your system unbootable. Take a make_tape_recovery backup before you start.

HP has a more detailed procedure in the priviledged document.

Maybe in here:
http://search.hp.com/redirect.html?url=http%3A//forums1.itrc.hp.com/service/forums/questionanswer.do%3FthreadId%3D475027&qt=ioinit+%2B-c&hit=2

SEP

[KapilRaj]

Multipathed tape devices ??

Kaps

[D Block 2]

Are there 4-paths to this device ?

TGT/LUN Device WWLUN_ID H/W_Path #_Paths
0/ 1 c12t0d1 6005-08B4-0010-041F-0000-9000-0032-0000 4

Are there Stand-by devices also?

It might appear that the hpux host can see multiple paths from the switch into the tape devices.

[Martin Corner]

Hi Again,

Thanks for everyones help so far.

Unfortunately nothing has changed. Zoning looks ok, we have an identical set-up in another site and that works ok. FC cards all at latest patches. Risked the ioinit -C command (couldn't do recovery tape as the MSL6030 is the only tape drive connected to server) but that just created exactly the same as before.

Output of fcmsutil /dev/td0 get remote all gives:

arget N_Port_id is = 0x450000
Target state = DVS_UNOPENED
Symbolic Name =
Port Type = N_PORT
Target Port World Wide Name = 0x100000e00222c166
Target Node World Wide Name = 0x100000e00202c166

Common Service parameters (all values shown in hex):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Version_Hi: 20 Version_Low: 20
BB_Credit: 2 Ciro: 1
Alt_BBcred: 1 Rxsz: 800
Conseq: ff RO_Bitmap: fe
Resolution: 0 EDTOV(ms): 7d0

Vendor version (in hex) 0 0 0 0
-------------------------------------------------------------


Target N_Port_id is = 0x450100
Target state = DVS_UNOPENED
Symbolic Name =
Port Type = N_PORT
Target Port World Wide Name = 0x10000000c93d10a0
Target Node World Wide Name = 0x20000000c93d10a0

Common Service parameters (all values shown in hex):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Version_Hi: 9 Version_Low: 8
BB_Credit: f Ciro: 1
Alt_BBcred: 0 Rxsz: 800
Conseq: ff RO_Bitmap: ffff
Resolution: 0 EDTOV(ms): 7d0

Vendor version (in hex) 0 0 0 0
-------------------------------------------------------------


Target N_Port_id is = 0x450101
Target state = DVS_READY
Symbolic Name = HSV100 - NC-EVA-0 - CR0135
Port Type = N_PORT
Target Port World Wide Name = 0x50001fe150035629
Target Node World Wide Name = 0x50001fe150035620

Common Service parameters (all values shown in hex):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Version_Hi: 9 Version_Low: 9
BB_Credit: 0 Ciro: 1
Alt_BBcred: 1 Rxsz: 800
Conseq: ff RO_Bitmap: 1
Resolution: 0 EDTOV(ms): 0

Vendor version (in hex) 0 0 0 0
-------------------------------------------------------------


Target N_Port_id is = 0x450102
Target state = DVS_READY
Symbolic Name = HSV100 - NC-EVA-0 - CR0135
Port Type = N_PORT
Target Port World Wide Name = 0x50001fe15003562d
Target Node World Wide Name = 0x50001fe150035620

Common Service parameters (all values shown in hex):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Version_Hi: 9 Version_Low: 9
BB_Credit: 0 Ciro: 1
Alt_BBcred: 1 Rxsz: 800
Conseq: ff RO_Bitmap: 1
Resolution: 0 EDTOV(ms): 0

Vendor version (in hex) 0 0 0 0
-------------------------------------------------------------

TIA

[Fabio Ettore]

Hi Sue,

I saw this problem at least 2 times. Problem is on configuration of Cisco MDS switches: robot of MSL9000 has to be on LUN0 and other drives from LUN1 on. It is possible that it is not so now for you.
Change it and then launch again

# ioscan -fnC tape

You should only see 2 tape drives.

Best regards,
Ettore

[Ashwani Kashyap]

fcmsutil queried only two remote tape targets .
One other thing I noticed is that your robotics appears on all the path . Robtics should only go with LUN0 .

HPUX cannot see anything unless it is asked to via zoning . I still doubt something is there in the cisco switch .

If you remove all the hardware paths of the tapes and the robotics using rmsf -H and if they come back after you do insf -e , then there definitley are multiple paths to the drives .

[Sanjay_6]

Hi,

Check and see how many of them are shown as claimed and how many of them are seen as NO_HW.

you may have to do some cleanup for the hardware paths / devices that no longer exist.

Hope this helps.

Regds

[Ray Carlson]

This has to do with how the tape drives are presented. Somelike like lun/target instead of target/lun. Then all the device files have to be removed and the HP re-booted. This should give you a starting point and I will try to check for the correct details on Monday.

[Martin Corner]

Hi,

Thank you so much for everyone who replied. The answer was a combination of all your answers plus a small config change on the ports on the Network Storage Router E1200.

So, on the E1200 port config of of FC Port 0 it was set to manual discovery only and default map was set to indexed. Once these were changed to Discovery on Link Up and Auto Assigned then we only had 1 autoch and 2 tapes claimed.

So then did the rmsf -H to remove the old hardware paths as Ashwani said.

Next step was to do a rmsf /dev/rmt/* and rmsf /dev/rac/* to get rid of the old device files.

Then did the mv ioconfig ioconfig.sav and the ioinit -c (no reboot) as recomended by Steven.

Then an insf -e -C tape -C autoch gives us the correct device files in /dev/rac and in /dev/rmt as was recomended.

Now everything work great.

Thanks again everyone, very much appreciated.