External DLT-8000 not seen on Tru64 Unix 5.1B (AS 4100)

[Horacio Rivero]

Hi,

Please help me with this problem. I connected an external DLT8000 tape drive on Alpha Server 4100 with Tru64 5.1B and SN-KZPCA-AX. I see the tape drive with ">>>show dev", but drive is not seen on Tru64.

- Is there a procedure to configure on Tru64 Unix 5.1B for the drive to work properly?
- is the drive supported on Tru64 5.1B?

Thank you

Regards

[Rob Leadbeater]

Hi,

Welcome to the ITRC forums.

That drive should be supported on 5.1B.

When you say that you can't see it in Tru64 what exactly are you doing to look for it ?

Can you see it if you do a

# hwmgr view devices

or

# hwmgr show scsi

Hope this helps,

Regards,

Rob

[Horacio Rivero]

Hi Rob,

thank you very much for your prompt response.

In fact when I do both "hwmgr view dev" and "hwmgr show scsi", I don´t see it.

[Ivan Ferreira · Accepted solution]

Maybe is connected to a scsi adapter that was not configured in the kernel yet, maybe a new one.

Try booting the generic kernel and compiling a new one:

boot -fi genvmunix -fl s

bcheckrc
cp /vmunix /vmunix.bak
doconfig
copy the new generated kernel to /
reboot

Run hwmgr to check if the device is detected.

[Mark Poeschl_2]

I suspect you're missing an entry in your /sys/conf/SYSTEMNAME file for the KZPCA. I don't recall exactly what name that device required, but there should be a line in the file that's something like 'config_driver itpsa' or 'config_driver aha_chim'. Boot from the generic kernel you should see the new tape drive. Then run 'doconfig' and you will get the appropriate drivers built in to the kernel.

[Horacio Rivero]

Hi,

thank you very much for your answers and help.

- What I will do is to copy new kernel to root with just "cp /newkernel /"?
- Is "doconfig" as is or with -c HOSTNAME or something else?

Any other suggestions?

thank you

Regards

[Horacio Rivero]

Hi,

thank you very much for your answers and help.

- What I will do is to copy new kernel to root with just "cp /newkernel /"?
- Is "doconfig" as is or with -c HOSTNAME or something else?

- What problems could ocur when doing doconfig?

Any other suggestions?

thank you

Regards

[Steven Schweda]

> - What I will do is to copy new kernel to
> root with just "cp /newkernel /"?

More like "cp -p /sys/NAME/vmunix /vmunix",
eventually, if not immediately. "doconfig"
should say something like:

The new kernel is /sys/NAME/vmunix

which should tell you where to find it.

> - Is "doconfig" as is or with -c HOSTNAME
> or something else?

That depends on which configuration file you
wish to use. Look around in
"/usr/sys/conf".

> - What problems could ocur when doing
> doconfig?

_Many_ problems _could_ occur when doing
"doconfig". Most are not likely, and if you
save the old "/vmunix", there's a limit to
how much damage you can do.


> Any other suggestions?

"man doconfig"

http://h30097.www3.hp.com/
http://h30097.www3.hp.com/docs/
http://h30097.www3.hp.com/docs/pub_page/doc_list.html
http://h30097.www3.hp.com/docs/pub_page/V51B_DOCS/V51B_DOCLIST.HTM
http://h30097.www3.hp.com/docs/base_doc/DOCUMENTATION/V51B_HTML/ARH9FETE/TITLE.HTM
http://h30097.www3.hp.com/docs/base_doc/DOCUMENTATION/V51B_HTML/ARH9FETE/CNFGMNTC.HTM
http://h30097.www3.hp.com/docs/base_doc/DOCUMENTATION/V51B_HTML/ARH9FETE/CNFGMNTC.HTM#static-config-sec
http://h30097.www3.hp.com/docs/base_doc/DOCUMENTATION/V51B_HTML/ARH9FETE/CNFGMNTC.HTM#add-device-config-sec

[Horacio Rivero]

Hi,

I am sorry for the questions but, I do not want to make mistakes.


- How do I know what configuration file to use?
- If something wrong happens, how do I restore the original kernel?

I am sorry again dear friends, can you tell me please how to do it, step by step ?

Thank you very much

Regards

[Ivan Ferreira]

- How do I know what configuration file to use?

You can create a new one by indicating no to "do you have an existing configuration file?", then you can select "All kernel options" when prompted.

Normally the configuration file is the hostname, so, when prompted by "do you have an existing configuration file", if you say yes, then you can accept the default prompted.

- If something wrong happens, how do I restore the original kernel?

Copy the /vmunix.bak (that you previously created with cp /vmunix /vmunix.bak) to /vmunix

I am sorry again dear friends, can you tell me please how to do it, step by step ?

Thank you very much

Regards

[Horacio Rivero]

Hi,

Ivan, thank you very much...now, what do you recommed, to use the default configuration file or the new one?
What are the "all kernel options" that you say and what to set?

Thank you again for your help

Regards

[Horacio Rivero]

So what I exactly will do is:

>>>boot -fi genvmunix -fl s
#bcheckrc
#cp /vmunix /vmunix.bak
#doconfig
.
.
.
Do you have an existing configuration file? N
.
.
.
The new kernel is /sys/HOSTNAME/vmunix
#cp /sys/HOSTNAME/vmunix /vmunix
#reboot
#hwmgr -view dev

Is that correct?
Is there some step wrong?

Should the system work fine when using this new configuration file?

thank you for all yor help

Regards

[Horacio Rivero]

Hi,

complementing my last reply, if something wrong happens, do I have to boot from "genvmunix" and the copy the saved "vmunix"?

I mean:

>>>boot -fi genvmunix -fl s
#cp /vmunix.bak /vmunix
#shutdown -r now

thank you

[Ivan Ferreira]

Your previous 2 posts are correct. That is the procedure to follow to create a new kernel and to rollback if something goes wrong.

[Horacio Rivero]

Dear friends,

I rebuilt kernel and tape drive is already seen with "hwmgr view dev", however I launched a "vdump" and aparently it was working good but did not finish because of an I/O error.

There were error messages before rebuild and error messages after rebuild........of course when booting (these were taken from messages file)

BEFORE REBUILD

Apr 23 19:32:29 ALPHA vmunix: itpsa0 at pci0 slot 3
Apr 23 19:32:29 ALPHA vmunix: IntraServer ROM Version V2.0 (c)1998
Apr 23 19:32:29 ALPHA vmunix: contig_malloc: cannot find memory in saved list
Apr 23 19:32:29 ALPHA vmunix: ITPSA: can't allocate script memory
Apr 23 19:32:29 ALPHA vmunix: hba_dme_attach: (3/n/n) - hba 'itpsa' attach failed.
Apr 23 19:32:29 ALPHA vmunix: cam_logger: SCSI event packet
Apr 23 19:32:29 ALPHA vmunix: cam_logger: bus 3
Apr 23 19:32:30 ALPHA vmunix: ccfg_simattach()
Apr 23 19:32:30 ALPHA vmunix: SIM Attach Failed, sim_init() reported FAILURE - Retry.
Apr 23 19:32:30 ALPHA vmunix: itpsa_attach_instance: Failure status returned from xpt_bus_register()
Apr 23 19:32:30 ALPHA vmunix: itpsa_confl1: itpsa0 attach failed


AFTER REBUILD

Apr 26 19:22:24 ALPHA vmunix: itpsa0 at pci0 slot 3
Apr 26 19:22:24 ALPHA vmunix: IntraServer ROM Version V2.0 (c)1998
Apr 26 19:22:24 ALPHA vmunix: scsi3: SCSI Error for target 5 sist0 = 0x1
Apr 26 19:22:24 ALPHA vmunix: cam_logger: SCSI event packet
Apr 26 19:22:24 ALPHA vmunix: cam_logger: bus 3
Apr 26 19:22:24 ALPHA vmunix: itpsa SCSI HBA
Apr 26 19:22:25 ALPHA vmunix: SCSI Error for target 5 sist0 = 0x1
Apr 26 19:22:25 ALPHA vmunix:
Apr 26 19:22:25 ALPHA vmunix: cam_logger: SCSI event packet
Apr 26 19:22:25 ALPHA vmunix: cam_logger: bus 3 target 5 lun 0
Apr 26 19:22:25 ALPHA vmunix: ss_perform_timeout
Apr 26 19:22:25 ALPHA vmunix: timeout on disconnected request
Apr 26 19:22:25 ALPHA vmunix: Active CCB at time of error
Apr 26 19:22:25 ALPHA vmunix: scsi3: HTH intr. on bus 3, SBCL = 0xa1
Apr 26 19:22:25 ALPHA vmunix: cam_logger: SCSI event packet
Apr 26 19:22:25 ALPHA vmunix: cam_logger: bus 3
Apr 26 19:22:25 ALPHA vmunix: itpsa SCSI HBA
Apr 26 19:22:25 ALPHA vmunix: HTH intr. on bus 3, SBCL = 0xa1
Apr 26 19:22:25 ALPHA vmunix:
Apr 26 19:22:25 ALPHA vmunix: scsi3: SCSI Bus was reset
Apr 26 19:22:25 ALPHA vmunix: cam_logger: SCSI event packet
Apr 26 19:22:25 ALPHA vmunix: cam_logger: bus 3
Apr 26 19:22:25 ALPHA vmunix: itpsa SCSI HBA
Apr 26 19:22:25 ALPHA vmunix: SCSI Bus was reset


What could be the problem?
Do you think is a faulty controller? or
Do yo think is a patch problem?


thank you.

[Rob Leadbeater]

Hi,

Is the SCSI bus on the tape drive properly terminated ?

That might be causing the SCSI errors during a vdump...

Regards,

Rob

[Horacio Rivero]

Hi,

Rob

There were SCSI error messages before rebuild and SCSI error messages after rebuild, but when booting system.

I did not get SCSI errors while doing "vdump", I just got an I/O error.

[Horacio Rivero]

I am sorry. About your question.....yes, the drive is properly terminated.

thank you

[Ivan Ferreira]

What is the device that you used as vdump destination? Use the device that ends in "c", for example:

/dev/tape/tape0c
/dev/ntape/tape0c

Can you post the output of the error?

You can also download LT&T to test the device and even generate a support ticket.

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareIndex.jsp?lang=en&cc=us&prodNameId=406731&prodTypeId=12169&prodSeriesId=406729&swLang=13&taskId=135&swEnvOID=1048

[Horacio Rivero]

Hi Ivan,

thank you very much for you help in my last trouble, it solved that problem. Now what I am really worry are the SCSI error messages when system is booting. (NOTE: this is after new kernel)

Apr 26 19:22:24 ALPHA vmunix: itpsa0 at pci0 slot 3
Apr 26 19:22:24 ALPHA vmunix: IntraServer ROM Version V2.0 (c)1998
Apr 26 19:22:24 ALPHA vmunix: scsi3: SCSI Error for target 5 sist0 = 0x1
Apr 26 19:22:24 ALPHA vmunix: cam_logger: SCSI event packet
Apr 26 19:22:24 ALPHA vmunix: cam_logger: bus 3
Apr 26 19:22:24 ALPHA vmunix: itpsa SCSI HBA
Apr 26 19:22:25 ALPHA vmunix: SCSI Error for target 5 sist0 = 0x1
Apr 26 19:22:25 ALPHA vmunix:
Apr 26 19:22:25 ALPHA vmunix: cam_logger: SCSI event packet
Apr 26 19:22:25 ALPHA vmunix: cam_logger: bus 3 target 5 lun 0
Apr 26 19:22:25 ALPHA vmunix: ss_perform_timeout
Apr 26 19:22:25 ALPHA vmunix: timeout on disconnected request
Apr 26 19:22:25 ALPHA vmunix: Active CCB at time of error
Apr 26 19:22:25 ALPHA vmunix: scsi3: HTH intr. on bus 3, SBCL = 0xa1
Apr 26 19:22:25 ALPHA vmunix: cam_logger: SCSI event packet
Apr 26 19:22:25 ALPHA vmunix: cam_logger: bus 3
Apr 26 19:22:25 ALPHA vmunix: itpsa SCSI HBA
Apr 26 19:22:25 ALPHA vmunix: HTH intr. on bus 3, SBCL = 0xa1
Apr 26 19:22:25 ALPHA vmunix:
Apr 26 19:22:25 ALPHA vmunix: scsi3: SCSI Bus was reset
Apr 26 19:22:25 ALPHA vmunix: cam_logger: SCSI event packet
Apr 26 19:22:25 ALPHA vmunix: cam_logger: bus 3
Apr 26 19:22:25 ALPHA vmunix: itpsa SCSI HBA
Apr 26 19:22:25 ALPHA vmunix: SCSI Bus was reset



There where also few SCSI errors when booting the system before new kernel and here the tape drive was not seen, so I do not what could be the problem.

These are the errors before new kernel:

Apr 23 19:32:29 ALPHA vmunix: itpsa0 at pci0 slot 3
Apr 23 19:32:29 ALPHA vmunix: IntraServer ROM Version V2.0 (c)1998
Apr 23 19:32:29 ALPHA vmunix: contig_malloc: cannot find memory in saved list
Apr 23 19:32:29 ALPHA vmunix: ITPSA: can't allocate script memory
Apr 23 19:32:29 ALPHA vmunix: hba_dme_attach: (3/n/n) - hba 'itpsa' attach failed.
Apr 23 19:32:29 ALPHA vmunix: cam_logger: SCSI event packet
Apr 23 19:32:29 ALPHA vmunix: cam_logger: bus 3
Apr 23 19:32:30 ALPHA vmunix: ccfg_simattach()
Apr 23 19:32:30 ALPHA vmunix: SIM Attach Failed, sim_init() reported FAILURE - Retry.

[Rob Leadbeater]

Hi,

OK. An I/O error during a vdump could be down to a number of things... A poor tape or a tape drive that needs cleaning are a couple of examples.

The SCSI errors during boot may be completely unrelated.

Have a look through the output of

# hwmgr show scsi -full

and see whether the scsi bus and lun mentioned in the errors are relevant to the tape drive.

Hope this helps,

Regards,

Rob

[Ivan Ferreira]

I cannot warranty that there is no problem with your SCSI device, but some SCSI bus reset messages can be considered as normal, even in this guide:

http://h30097.www3.hp.com/docs/base_doc/DOCUMENTATION/V51B_ACRO_DUX/ARHGXETE.PDF

You can see the sample output of a Tru64 boot with bus reset messages.

[Horacio Rivero]

Hi,

Is it possible that the ploblem is caused by SCSI controller or a missing patch or something else?

I think the Administrator is not convinced that the messages are normal.

Thank you

[DCBrown]

SCSI Error for target 5 sist0 = 0x1

sist0 is set when a bus parity error is detected according to the documentation I found on the web for Symbios.

HTH intr. on bus 3, SBCL = 0xa1

This message is output when the driver believes the board is wedged and needs to be reset. This may be related to the original bus parity error.

Just a guess, but I'd say there is still a h/w issue somewhere. Termination and cable or cable length issues are the prime suspects. There shouldn't be scsi bus parity errors occurring.

[Horacio Rivero]

Hi,

Thank you DCBrown, what I will do is to verify cable and termination again, also replace SCSI controller.

I will give my feedback.

Regards