Unsolved

This post is more than 5 years old

9 Posts

2395

September 19th, 2006 04:00

Inquire commnad cannot show tape drives

Hi,

We had a QLogic HBA in a Solaris 8 Storage Node.

We have a ACSLS Silo with a STK SL8500 library in a SAN environment. This library has several IBM LTO2 FC configured as DDS.

With QLogic HBA, the inquire command could see the DDS Tape Drives.

Now we have install a new EMC CX700 Storage and we have replace the QLogic HBA with a LP9802 HBA. We have also installed in the SN host EMC PowerPath.
We have created the persistent binding in HBA (/kernel/drv/lpfc.conf). We have recreated the drive paths (/dev/rmt/xxcbn) without problems.

The problem is that now the command inquire cannot see the DSS Tape Drives. It only can see the disks of CX700 Storage, but O.S. can see the tapes.

We have other SN that are sharing this Tapes drives without problems. I need the information of the command inquire to re-run jbconfig with ACSLS Silo (I need the path (dev/rmt/xxcbn and the WWNN od the tape drive).

Why inquire command cannot see the Tape Drives? Any idea?
How can I know the WWNN of each path /dev/rmt/xcbn?

Thanks

9 Posts

September 19th, 2006 04:00

Sorry!, Networker version in SN and server is 7.1.3

Thanks

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

September 19th, 2006 05:00

If OS can see the drives and inquire can't I would suspect lus.conf. I'm currently at site where environment is like yours apart using Emulex 10k HBAs. I had to configure lus.conf to see drives (10k was missing there so I had to define it and add range inquire will scan). I believe that is your issue too.

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

September 19th, 2006 07:00

If I'm right there is a big chance that in /var/adm/messages upon startup of the machine you will find message saying "no match for found in lus.conf"

9 Posts

September 19th, 2006 07:00

Hi,

What must I configure in lus.conf????

Thanks

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

September 19th, 2006 07:00

Probably your hba and range. Use lusbinfo to scan buses and see how your hba gets reported. Then check lus.conf for it and if it is not defined then define it as others are defined. Once that is done reload lus module.

9 Posts

September 19th, 2006 08:00

If I'm right there is a big chance that in
/var/adm/messages upon startup of the machine you
will find message saying "no match for found
in lus.conf"


Hi,
You are right! In /var/adm/messages there is this message:

**********************************
Sep 15 15:42:39 ejsun05 scsi: [ID 193665 kern.info] lus8 at lpfc0: target 0 lun 0
Sep 15 15:42:39 ejsun05 genunix: [ID 936769 kern.info] lus8 is /pci@1f,2000/lpfc@1/lus@0,0
Sep 15 15:42:40 ejsun05 lus: [ID 511245 kern.info] no match for pci10df,f980 found in lus.conf:
Sep 15 15:42:40 ejsun05 lus: [ID 321757 kern.info] ntargets - 8 - based on bus width
Sep 15 15:42:40 ejsun05 scsi: [ID 193665 kern.info] lus9 at lpfc1: target 0 lun 0
Sep 15 15:42:40 ejsun05 genunix: [ID 936769 kern.info] lus9 is /pci@4,2000/lpfc@1/lus@0,0
Sep 15 15:42:40 ejsun05 genunix: [ID 115966 kern.warning] WARNING: add_spec: No major number for mpt
****************************
I have two HBA; lpfc0 and lpfc1

The hba_id is pci10df,f980

My /usr/kernel/drv/lus.conf is:
---------------------------------------------
known-scsi-adapters="esp:8-7 SUNW,esp:8-7 isp:16-7 SUNW,isp:16-7 QLGC,isp:16-7 fas:16-7 SUNW,fas:16-7 glm:16-7 SUNW,glm:16-7 pci1000,f:16-7 ifp:126-126 SUNW,ifp:126-126 PTIX,isptwo:126-126 pci1077,2100:126-126 pci1077,1:126-126 pci9004,8478:16-7 pci10df,f700:126-32C pci10df,f800:126-32C pci10df,f900:126-32C fca:126-32C fcaw:126-32C pci1242,4643:126-32C pci1077,2200:126-32 JNI,FCE:126-32C fce:126-32C JNI,FCR:126-32C fcr:126-32C fp:126-32 pci1077,2:126-32 pci1077,9:126-32 pci10cf,105c:126-32C pci10cf,1178:126-32C fjulsa:16-7 ";

name="lus" class="scsi" target=0 lun=0;
-------------------------------------------------------------------
Must I add the string to known-scsi-adapters section???
Must I reboot the machine?????

Thanks

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

September 19th, 2006 09:00

Hello,

Yes, you should add that. To be 100% sure at the bottom you can add:

name="lus" class="scsi" target=0 lun=0;
name="lus" class="scsi" target=1 lun=0;
name="lus" class="scsi" target=2 lun=0;
name="lus" class="scsi" target=3 lun=0;
name="lus" class="scsi" target=4 lun=0;
name="lus" class="scsi" target=5 lun=0;
name="lus" class="scsi" target=6 lun=0;
name="lus" class="scsi" target=7 lun=0;
name="lus" class="scsi" target=8 lun=0;
name="lus" class="scsi" target=9 lun=0;
name="lus" class="scsi" target=10 lun=0;
name="lus" class="scsi" target=11 lun=0;
name="lus" class="scsi" target=12 lun=0;
name="lus" class="scsi" target=13 lun=0;
name="lus" class="scsi" target=14 lun=0;
name="lus" class="scsi" target=15 lun=0;
name="lus" class="scsi" target=16 lun=0;
name="lus" class="scsi" target=17 lun=0;
name="lus" class="scsi" target=18 lun=0;
name="lus" class="scsi" target=19 lun=0;
name="lus" class="scsi" target=20 lun=0;


If you want you can reboot machine, but modunload (see pid with modinfo) and then modload of lux will avoid reboot option.

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

September 19th, 2006 09:00

update_drv will do reload (modunload+modload) - not sure if you have it on Solaris 8 (I used it under 9).

9 Posts

September 25th, 2006 04:00

Hi, again:

I have added the entry pci10df,f980:126-32C and the end of known-scsi-adapters section,but:

a) Before add this entry, inquire takes about 3-4 minutes and it can not see the tapes. The last device it can see is a disk in scsidev@5.7.0

b) After add this entry, inquire takes about 2 hours, but it can see the tapes:
scsidev@2.10.0:IBM ULTRIUM-TD2 Tape /dev/rmt/1cbn
scsidev@5.10.0: IBM ULTRIUM-TD2 Tape /dev/rmt/2cbn

Now, inquire command search for lun 0-31 in each target, and for many targets in each scsi ID (I killed the command in scsidev@5.45.31)

I think that entry pci10df,f980:126-32C is not correct but the problem is lus.conf

The ouput of lusbinfo commnad is:

Bus #2 is 'pci10df,f9800', dma_max=-1, initiator ID=255, ntargets=8, nluns=7
Bus #5 is 'pci10df,f9801', dma_max=-1, initiator ID=255, ntargets=8, nluns=7

What is the correct entry I must add in lus.conf in order to tape drives can see detected but the commnad inquire runs in a few minutes???

Thanks

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

September 25th, 2006 05:00

You can restrict number of luns inspected with range you configure (read comments in lus.conf for more info). There is also env variable for that I believe and that information should be somewhere in docs.

I don't know on which bus is your hw so I can't tell you which one to use. From output you sent it looks like you use both so add both. Also there are few switches that might speed up the things with inquire, but they are not your issue here.

79 Posts

November 6th, 2007 10:00

Hi,

I too have a similar issue on solaris 9 Networker server ( 7.3.3) . It has 2 LP9802 HBA's connected to 2 Fibre-SCSI bridges and they are connected to ATL P7000 SCSI Tape Library. What's happening is we have Solaris /dev/rmt/ showing all the 16 Tape drives but "inquire" is showing only 6 Tape drives on each path.

I suppose I have to update lus.cong but since we have 2 paths I wanted to know what entries should I have to add .

Thanks in advance.

79 Posts

November 6th, 2007 10:00

Hi ,

I would like to know finally what is the resolution you found for this probelm.
Because we too have simialr problem on one of your networker server


Thanks in advance.

6 Operator

 • 

14.4K Posts

 • 

56.2K Points

November 6th, 2007 23:00

Add your card and range for it as described above. Use lusbinfo command to check what HBA you use (you can also check /var/adm/messages for lus messages as they would appear upon OS booting up in case that card is not part of your existing lus.conf).

79 Posts

November 12th, 2007 07:00

Thank you so much. This solved my problem.
No Events found!

Top