JNIC v5.3.0.1

[jjrinconbayon]

Hi,

i have a JNI SCSI Fibre Channel host bus adapter in a Sun Fire 480R, and it is working properly with a driver version 5.2.3. We want to upgrade this driver to the version 5.3.0.1, but when we do it, the Operating System can see all drives (mt -f /dev/rmt/1cbn status) but Legato Networker can´t do it (/etc/LGTOuscsi/inquire).
We have statis binding by Port-Id.

do you have any suggestions?,

Regards,
Juanjo

 

[wallace88]

According to the SAN Component Compatibility Guide:
(

http://portal1.legato.com/resources/compatibility/lanfree.cfm),

The following is the compatiability matrix for Sun and JNI HBA's:

Sun Solaris 8/9 JNI FCI-1063 PCI 2.5
Sun Solaris 8/9 JNI FC64-1063 S-BUS 13.3.7
Sun Solaris 8/9 JNI FCE-1473, FCE2-1473 S-BUS v3.89
Sun Solaris 8/9 JNI FCE-6460*** PCI (Driver v5.1.1)
Sun Solaris 8/9 JNI FCX-6562(-L), FCX2-6562(-L) PCI (Driver: 1.0b14)


The matrix shows what HBA's have been tested and certified to work.


Isn't the simple answer to go back to the driver version that works?

 

[julianbarnett]

You need to make sure the /usr/kernel/drv/lus.conf is updated with the correct details. Has the name of the driver changed when you look at the /var/adm/messages file during boot? Also are the devices still on the same scsi path - sometimes a reconfigure reboot will bring devices up on a different channel if boards have been added to the server. The lus.conf file describes the format of the entries. You can reload the lus driver in flight by calling..

cd /usr/kernel/drv
modinfo | grep lus -- First field is the driver id
modunload -i <id> <-- from above
modload lus

regards Julian

 

[jjrinconbayon]

it seems that the driver is loaded properly

the output of the command 'modinfo | grep -i jni' is
131 ..... jnic146x (JNIC v5.3.0.1 (03080800))

and if i remove the path devices to the drives and i generate them again the Operatin System see the drives in the same paths

rm /dev/rmt/*
devfsadm -v

or

rm /dev/rmt/*
touch /reconfigure
boot

 

[julianbarnett]

The jni driver is not the issue, it's the LUS driver - this is what Legato inquire command uses to detect drives/robots. If mt commands can see the drives, rewind, status etc. but inquire can't then lus.conf seems to be at fault. We recently upgraded out HBA cards and found that the default lus driver didn't recognise them - had to add an entry in lus.conf with the driver name as reported by the kernel and the addresses it should recognise i.e.

known-scsi-adapters=&quot;.....<snip>... lpfs:126-32C &quot;;

name=&quot;lus&quot; class=&quot;scsi&quot; target=30 lun=0 hba=&quot;lpfc0&quot; wwn=100000e00221c53c;
name=&quot;lus&quot; class=&quot;scsi&quot; target=35 lun=0 hba=&quot;lpfc2&quot; wwn=500308c00145d80e;

The name lpfs is the name the name of the HBA device e.g.

ls -l /dev/rmt/11cbn
lrwxrwxrwx 1 root root 43 Nov 13 17:32 /dev/rmt/11cbn - ../../devices/sbus@2,0/lpfs@1,0/st@22,2:cbn

Can you post the output of your /usr/kernel/drv/lus.conf file excluding any blank lines or comments, and an ls -l of one of your tape devices?

 

[jjrinconbayon]

It could be what you say 'julianbarnett', i'm going to change out lus.conf with the entry 'JNI,FCR:126-32C', as soom as possible
My question is why with the other drive version is working
Thanks again,
Juanjo

#cat /usr/kernel/drv/ls.conf

known-scsi-adapters=&quot;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 fp:126-32 pci1077,2:126-32 pci1077,9:126-32 pci10cf,105c:126-32 &quot;;

name=&quot;lus&quot; class=&quot;scsi&quot; target=0 lun=0;


# ls -la /dev/rmt/1cbn
lrwxrwxrwx 1 root root 47 Nov 27 16:55 /dev/rmt/1cbn -> ../../devices/pci@8,600000/JNI,FCR@2/st@1,0:cbn

 

[jjrinconbayon]

wallace88, the driver version v5.3.0.1 can be upgraded without having to reboot the server.
And the Sun Microsystem service support assure me that this version is certified to work

Thanks again,
Juanjo

 

[wallace88]

Did SUN certify that v5.3.0.1 will work with NetWoker?

 

[jjrinconbayon]

yes this is what they told me, ...

 

[wallace88]

My point is this: If Legato's compatiability matrix does not show that your particular HBA and driver version is supported, then you cannot expected Legato for help if it does not work.

The reason is that, to my knowledge, there is no one standard regarding HBA and fiber protocols. This is the reason why the compatiability matrix exists: to tell customers what HBA has been certified to work with what o/s platforms.

If SUN says it is supported, and you have support with SUN, then you have to ask them to resolve the issue.

Again, if things worked with the older driver, why not just use that driver?

As others suggested, verify that the lus.conf has been set up correctly. Worse comes to worse... reinstall the LGTOdrvr module. (this means uninstalling LGTOserv and LGTOnode before hand, then reinstalling these modules as well).

 

[julianbarnett]

Maybe the driver has changed name - was the previous driver reporting JNI,FCE? which was already in the lus.conf file. I know our &quot;supported&quot; HBA cards LP9002S did not have an entry for their driver name (lpfs) in the lus.conf and until I added it and reloaded the lus module we saw exactly the same symptoms as you.

 

[jjrinconbayon]

i added the entry JNI,FCR:126-32C in the file lus.conf and i reloaded the module and is doesn`t work.
In the /var/adm/messages file i have the following line

... jnic146x0: ÌD 665907 kernel.notice] jnic146x0: Hba: JNI,FCR Model: FCE-6460-N

thanks again,
Juanjo

 

[LGTOserv]

JNI 5.3.x.x + has changed their driver. Also ; changed the name to FCR, you require alterations to the JNIxxx.conf and a new lus.conf. Contact Legato for this...