Disk hardware path: I/O request failed

[Andrew Merritt_2]

Do you have any documentation that suggests this Sun disk array is supported on this HP system?

Andrew

[Steve Lewis]

The fact that the error message said that the command is not supported by the device tells me that your SUN firmware disks are not fully compatible with the HP-UX kernel's SCSI commands.

Have you tried /usr/contrib/bin/scsi_io ?

If there is no controller inbetween the individual disks and the HBA, then you could try a test firmware write of HP firmware onto a disk, but only if you don't mind the likely outcome of a toasted disk.

My gut feeling is that even though its all SCSI, it just won't work. Plug the array back into a SUN box and use it through NFS or Samba.

[Robert-Jan Goossens]

Hi Yogeeraj,

Could you attach this enclosure to a SUN server ?

# format
select a disk
enter i (for the inquire command)
q (quit)

it should be revision 0828 otherwise you need to download below patch.

108105-02

Hope this helps,
Robert-Jan

[Robert-Jan Goossens]

Sorry my mistake latest firmware rev is 6329

same patch.

[Andrew Merritt_2]

Steve,
If by "The fact that the error message said that the command is not supported by the device" you're referring to:
"The selected device is not supported by this tool.
Inaccurate information could be displayed by the tool.",

it's not quite the same thing. That message is saying the STM info tool does not recognise the device since it's not using HP firmware. However, in all probability that does also imply that the device is not supported on the HP hardware, since the STM disk info tool should support all supported SCSI disks.

It's a minor point, but I think worth making for clarity.

Andrew

[Yogeeraj_1]

dear all,

Feedback.

I did the reseating of the SCSI cables this morning but it is still giving me the same errors.

What are the probability of an impending disk failure?

I will have to fetch my SUN server from the "dungeons" to be able to test the storage box again ;)

I have the possibility to try its twin which is still untouched...

more views will be most appreciated...

best regards
yogeeraj

[Gerhard Roets]

Hi Yogee

Are you using inline terminated scsi cables? In that case make sure there is no "double" terminations on either end of the scsi bus.

Regards
Gerhard

[Dave Unverhau_1 · Accepted solution]

Hi Yogeeraj,

First, I'd like to clear something up... what you're trying to use is actually "just a bunch of disks" (JBOD), right? This doesn't look like an "array", based on the ioscan info...

Assuming that this is just a box of disk drives attached to an ultra-SCSI bus on the L1000, and nothing else is on that bus, (and assuming that I remember how to correctly decode the SCSI timeout error you provided), it looks like the drive at address 1 is not behaving well on the bus (according to this one message).

The line bp->b_dev: cb001000 indicates that the device in error is SCSI address 1, LUN 0, on bus 0. If you are getting a number of these SCSI timeout messages and the b_dev is always the same, that drive *probably* has a problem. Whether or not flashing the firmware will fix it is hard to tell.

Now, if you're seeing these messages with different b_dev numbers, the problem is probably bus cabling, termination, power, etc.

It's also possible that a cabling problem (bent/broken pin, failed terminator) could cause a problem to point to a single drive if it causes that drive to not be addressed correctly.

The first thing I'd do if the errors point to drive 1 is to disconnect it (power and SCSI connections is always best) and see how things change. You may need to (and should) shutdown and power off the system when making SCSI bus connection changes.

I hope this helps you get things straightened out!

Best Regards,

Dave

[Yogeeraj_1]

hi,

i have already tried to reseat the cables. still not ok!

could there be any chances of Disk Crash causing these errors to occur?

also, i have observed that the device errors are all in the 4 addresses
===============================
#grep -i "bp->b_dev" /var/adm/syslog/syslog.log
...
...
Feb 7 07:23:10 l1000 vmunix: bp->b_dev: cb001000
Feb 7 07:23:22 l1000 vmunix: bp->b_dev: cb001000
Feb 7 07:23:35 l1000 vmunix: bp->b_dev: cb002000
Feb 7 07:23:48 l1000 vmunix: bp->b_dev: cb003000
Feb 7 07:24:01 l1000 vmunix: bp->b_dev: cb004000
.....
=============================

more insight will help a lot.

regards
yogeeraj

[Yogeeraj_1]

Hi again,

We are still having problems troubleshooting this. The STM is not indicating any errors or warnings. But the syslog has been logging lots of the SCSI errors. (attached the latest syslog.log)

Could that be related to "incompatible" SCSI cable/card connections? How do i verify?

thanks in advance.

regards
yogeeraj

[Eric Antunes]

Hi Yogeeraj,

Check this thread (you may have 2 SCSI adapter on same BUS...):

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=609801

Finaly, if you can't get any results, check if you have the latest SCSI patches:

#swlist |grep SCSI

Best Regards,

Eric Antunes

[RAC_1]

I am running out of possible causes that might cause this. Would it be possible for you to run dd on each of the disks and check for error??
To me, the errors from EMS seems to be effect of the fact that this is SUN thing on hp-ux.

The other thing that you might want to have a look at is latest SCSI patches.

Anil

[Yogeeraj_1]

hi,

below the swlist concerning the SCSI patches, am not sure if this is OK.
===============================================
# swlist -l patch |grep -i scsi
# PHCO_25831 1.0 SCSI Ultra160 driver Online Addition script
# PHCO_29098 1.0 scsictl(1M) cumulative patch
# PHKL_23292 1.0 scsi_surface.h header file patch
# PHKL_23666 1.0 SCSI IO Subsystem Cumulative Patch
# PHKL_23810 1.0 Enable SCSI floppy for 64 bit computers
# PHKL_28209 1.0 INTRMIGR,fast boot enabler,iSCSI ioscan fix
# PHKL_29899 1.0 SCSI Tape (stape) cumulative
# PHKL_30511 1.0 SCSI Ultra160 CumulativePatch
# PHKL_30523 1.0 PDC_SCSI_PARMS,iCOD,vPars,IPMI,OLA/R
# PHKL_32090 1.0 SCSI IO Cumulative Patch
# SCSI-Passthru B.11.11 HP SCSI-Passthru Driver
# SCSI-Passthru.SPT-DVR B.11.11 NIO SCSI Pass Thru Driver (SPTD)
# SCSI-Passthru.SPT2-DVR B.11.11 Machine specific files for NIO SPTD
#
============================================

regards
Yogeeraj

[Yogeeraj_1]

hi,

my disks have been tested ok.

could this be a "SCSI problem"?
The Storage Box is ULTRA SCSI and the port on my L1000 is a SCSI.

experts please advise.

NB. i have also 2 Differential SCSI ports available on the server (Was initially installed for a AutoRAID 12H).

regards
yogeeraj

[Eric Antunes]

Hi Yogeeraj,

Can you post the full io?

#ioscan -fn > io_output.txt

Are you able to run vgdisplay on that volume group?

Best Regards,

Eric Antunes

[Yogeeraj_1]

hi eric,

attached the output of the ioscan -fn

not that i have created no volume groups. I will be using them as Disks in ASM to create my Disk Groups for Oracle 10g.

regards
yogeeraj

[Andrew Merritt_2]

You really need to contact the manufacturer of the device (Sun, in this case) to see if it is supported on HPUX. From the troubles you are having, I rather suspect not.

Andrew

[Eric Antunes]

Yogeeraj,

I think you do have 2 SCSI interfaces (like I doubted before when I gave you a link with a similar issue) on same BUS and that may be the cause of your troubles.

You have:

ext_bus 0 0/0/1/0 c720 CLAIMED INTERFACE SCSI C896 Fast Wide Single-Ended

and

ext_bus 1 0/0/1/1 c720 CLAIMED INTERFACE SCSI C896 Ultra Wide Single-Ended

You are using 0/0/1/... for both scsi interfaces.



I have:

ext_bus 1 0/0/1/1 c720 CLAIMED INTERFACE SCSI C896 Ultra Wide Single-Ended

and

ext_bus 2 0/0/2/0 c720 CLAIMED INTERFACE SCSI C875 Ultra Wide Single-Ended


I'm using 0/0/1/... for first and 0/0/2/... for second scsi interface.

Eric Antunes

[D Block 2]

Yogeeraj,

let's try this: disconnect the second cable to the scsi jbod and reboot (you have a single-ended scsi, I don't think you need a terminate on your server side).

what does the ioscan show ?

also, with ASM, did you change the owner ship of these devices to oracle:oinstall ?

note: if you run: "sam", this forces an insf -e and the ownership changes back to root.

[Yogeeraj_1]

hi Eric,

Am a little lost now! Do you mean that i have connected the cable to the wrong interface on the server?

..but i have only one (which is not Fast Wide Differential)!!

how do i know?

let me know.

NB. I have already documented myself about all ASM configuration procedures. So no problem on this side ;)

thanks
yogeeraj

[Yogeeraj_1]

hi again,

A summary of my current configuration for SCSI interfaces:

0/0/1/0 : external storaged box

0/0/1/1 : 2 Internal disks
0/0/2/0 : 2 Internal disks

0/0/2/1 : dvd rom drive

0/4/0/0 : FW SCSI port (reserved for autoRAID 12H)
0/4/0/1 : FW SCSI port (reserved for autoRAID 12H)

0/7/0/0 : FW SCSI port (reserved for autoRAID 12H)
0/7/0/1 : FW SCSI port (reserved for autoRAID 12H)

It seems that the card used for the external storage box is on the same bus as the Internal Disks and this may be causing a problem.

Should i move the first card to another slot on the server?

thank you for futher guidances


regards
yogeeraj

[Eric Antunes]

Hi Yogeeraj,

Sorry, I gave you an incomplete answer. Let's see if I explain me better...


All my system disks are connected to JUST one "SCSI C... Ultra Wide Single-Ended":

ext_bus 1 0/0/1/1 c720 CLAIMED INTERFACE SCSI C896 Ultra Wide Single-Ended
target 1 0/0/1/1.2 tgt CLAIMED DEVICE
disk 0 0/0/1/1.2.0 sdisk CLAIMED DEVICE SEAGATE ST39103LC
/dev/dsk/c1t2d0 /dev/rdsk/c1t2d0

AND


ext_bus 2 0/0/2/0 c720 CLAIMED INTERFACE SCSI C875 Ultra Wide Single-Ended
target 3 0/0/2/0.2 tgt CLAIMED DEVICE
disk 1 0/0/2/0.2.0 sdisk CLAIMED DEVICE SEAGATE ST39103LC
/dev/dsk/c2t2d0 /dev/rdsk/c2t2d0



And all my 6 apps and database disks are connected to 2 "SCSI C896 Ultra2 Wide LVD" BUT with help of the SureStore disk system "HP A5272A":


ext_bus 4 0/4/0/0 c720 CLAIMED INTERFACE SCSI C896 Ultra2 Wide LVD
target 7 0/4/0/0.7 tgt CLAIMED DEVICE
ctl 4 0/4/0/0.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c4t7d0
target 8 0/4/0/0.8 tgt CLAIMED DEVICE
disk 3 0/4/0/0.8.0 sdisk CLAIMED DEVICE SEAGATE ST136403LC
/dev/dsk/c4t8d0 /dev/rdsk/c4t8d0
target 9 0/4/0/0.9 tgt CLAIMED DEVICE
disk 4 0/4/0/0.9.0 sdisk CLAIMED DEVICE SEAGATE ST136403LC
/dev/dsk/c4t9d0 /dev/rdsk/c4t9d0
target 10 0/4/0/0.10 tgt CLAIMED DEVICE
disk 7 0/4/0/0.10.0 sdisk CLAIMED DEVICE SEAGATE ST136403LC
/dev/dsk/c4t10d0 /dev/rdsk/c4t10d0
target 11 0/4/0/0.14 tgt CLAIMED DEVICE
ctl 8 0/4/0/0.14.0 sctl CLAIMED DEVICE HP A5272A
/dev/rscsi/c4t14d0
ext_bus 5 0/4/0/1 c720 CLAIMED INTERFACE SCSI C896 Ultra2 Wide LVD
target 12 0/4/0/1.7 tgt CLAIMED DEVICE
ctl 5 0/4/0/1.7.0 sctl CLAIMED DEVICE Initiator
/dev/rscsi/c5t7d0
target 13 0/4/0/1.8 tgt CLAIMED DEVICE
disk 5 0/4/0/1.8.0 sdisk CLAIMED DEVICE SEAGATE ST136403LC
/dev/dsk/c5t8d0 /dev/rdsk/c5t8d0
target 14 0/4/0/1.9 tgt CLAIMED DEVICE
disk 6 0/4/0/1.9.0 sdisk CLAIMED DEVICE SEAGATE ST336704LC
/dev/dsk/c5t9d0 /dev/rdsk/c5t9d0
target 15 0/4/0/1.10 tgt CLAIMED DEVICE
disk 8 0/4/0/1.10.0 sdisk CLAIMED DEVICE SEAGATE ST136403LC
/dev/dsk/c5t10d0 /dev/rdsk/c5t10d0
target 16 0/4/0/1.14 tgt CLAIMED DEVICE
ctl 9 0/4/0/1.14.0 sctl CLAIMED DEVICE HP A5272A
/dev/rscsi/c5t14d0


SO, you may need 1 SureStore disk system device to connect several disks to the same SCSI interface. Neverthless, it's better for you to confirm this with an HP engineer...


Best Regards,

Eric Antunes

[Eric Antunes]

Hi Yogeeraj,

Did you solved your issue? If not, you may also consider to post it in the server forum:

http://forums1.itrc.hp.com/service/forums/familyhome.do?familyId=122

Best Regards,

Eric Antunes

[Yogeeraj_1]

hi Eric,

Problem is still pending.

In the meanwhile i have been able to get an autoRAID 12h tower and am working with it.

I have not found the solution yet.

I will try to post in the Servers section.

thanks again for your help
cheers
yogeeraj