Kernel panic after adding disks

[DD-CC]

Hi there

I'm having problem adding a extra diskarray (A3311A) with 6 new disks to my K220 server running hpux 10.20 on the F/W SCSI. After trying to load the vmunix kernel the kernel panics and the system crashes.
The last thing before the kernel panics was something about scsi id 10/8.7.0, but - I can't recall the message from memory.

I want to add 4 disks on scsi id 10/8.9.0, 10/8.10.0, 10/8.11.0 and 10/8.13.0 and 2 disks on 10/0.9.0 and 10/0.11.0.
I've checkt the scsi id settings on an other server by scanning the scsi bus and the addresses are setup correct. I've also checked the cables and they are ok.

Output from ioscan:

H/W Path Class Description
===============================================
bc
8 bc I/O Adapter
10 bc I/O Adapter
10/0 ext_bus GSC built-in Fast/Wide SCSI Interface
10/0.1 target
10/0.1.0 disk SEAGATE ST318436LC
10/0.2 target
10/0.2.0 disk SEAGATE ST39173WC
10/0.3 target
10/0.3.0 disk SEAGATE ST318436LC
10/0.4 target
10/0.4.0 disk SEAGATE ST318436LC
10/0.5 target
10/0.5.0 disk SEAGATE ST34572WC
10/0.7 target
10/0.7.0 ctl Initiator
10/0.12 target
10/0.12.0 disk SEAGATE ST318436LC
10/0.14 target
10/0.14.0 disk SEAGATE ST39173WC
10/0.15 target
10/0.15.0 disk SEAGATE ST39173WC
10/4 bc Bus Converter
10/4/0 tty MUX
10/4/16 lan HP HP-PB 100 Base TX card
10/8 ext_bus GSC add-on Fast/Wide SCSI Interface
10/8.1 target
10/8.1.0 disk SEAGATE ST318436LC
10/8.2 target
10/8.2.0 disk SEAGATE ST39173WC
10/8.3 target
10/8.3.0 disk SEAGATE ST318436LC
10/8.4 target
10/8.4.0 disk SEAGATE ST318436LC
10/8.5 target
10/8.5.0 disk SEAGATE ST34371W
10/8.7 target
10/8.7.0 ctl Initiator
10/8.12 target
10/8.12.0 disk SEAGATE ST318436LC
10/8.14 target
10/8.14.0 disk SEAGATE ST39173WC
10/8.15 target
10/8.15.0 disk SEAGATE ST39173WC
10/12 ba Core I/O Adapter
10/12/0 ext_bus Built-in Parallel Interface
10/12/5 ext_bus Built-in SCSI
10/12/5.0 target
10/12/5.0.0 tape HP C1537A
10/12/5.2 target
10/12/5.2.0 disk TOSHIBA CD-ROM XM-5401TA
10/12/5.7 target
10/12/5.7.0 ctl Initiator
10/12/6 lan Built-in LAN
10/12/7 ps2 Built-in Keyboard/Mouse
32 processor Processor
34 processor Processor
36 processor Processor
38 processor Processor
49 memory Memory


Any idear what I could do ??

Regards
S?ren Dreyer

[Insu Kim]

I guess that there's a problem associated with SCSI address conflict.

My suggestion is to stop the boot process to get into the PDC Main Menu after adding 6 disk drives.
And there, type search to make sure that system will see 6 different disks set to different SCSI IDs.
If this step goes well, the problem will be gone.

Good lunk,

[Tim Malnati]

Jamaica enclosures are pretty straight forward; very little can go wrong with the equipment, but I have seen situations where the dip switches have failed.

The first question is everything properly terminated and within length constraints. If this is ok, then I suggest that you attempt to startup the system with the new enclosure connected but the drives removed. If this evolution fails your problem is is probably a bad cable or terminator.

If the system comes up normally, you probably have a scsi id problem. I would now try installing disks individually. Typically you can do this hot, but where you are having panics I would suggest that you perform each from cold instead to prevent corrupting any files. The bad scsi id will become obvious using this method. You might be misreading the dip switch or it is truely bad. If it's bad though, make sure that it gets repaired now. You don't need a situation down the street where someone slips in a disk and causes your machine to take a hit.

[paul courry]

Another tool to use is ODE MAPPER. This will allow you to ping each SCSI address and see what is at this address. Multiple devices or something set to the controller address at 7 will show immediately and won't crash the system. This is a common tool used by CE's to diagnose hardware problems, you will find it on your support disk.

[DD-CC]

Hi, and thanks for the reply

The diskarray is an externally rack, and I'm trying to connect 4 disks to F/W scsi I/O on hw address 10/0 and 2 disks on F/W scsi I/O on hw address 10/8 - with two different sets of cables.

I've checkt the settings of the dipswitches by scanning the scsi bus on an other server, and they are set correct.

When I interact with the bootprocess and do a 'search' for bootable devices all new disks are displayed.

Now, the server is (of cause) an important server so I can't get my hands on it again for at least an other week. At that time I'll try the ODE tool to scan the scsi to see what's on the bus and test the disks.

The diskarray I'm using was running fine on an other server a couple of days ago, so it should be ok. If ODE dosen't show any errors, I'll try anyway to connect the diskarray without disks and boot the server, and then plug in disk by disk (and rebootding each time) to find the sinner-disk.

... or should I try something else ?

[Insu Kim]

You don't need to reboot the system after adding the disk drives step by step in the enclosure.
Just doing "ioscan -Cdisk -fn" will let HPUX recorgnize the disk drives installed.

There're two more things to be considered even if you verified cables, terminators and scsi length constraint.
One is the cable length which is daizy-chained for the purpose of using full slots in the enclosure with one scsi adaptor attached.
The recommedation is 0.5 meter.

The other is fuses misloaded on F/W scsi interface. there must be 5 amp fuses at the location F1, F2 on the adaptor but on some adaptor, 3/8 fuses are placed which can induce scsi problems.
If you're certain everything but problem is still there, place a hardware support call to HP to investigate these fuses on the adaptors.
These fuses can cause a problem in a scsi bus with a load.
In your case, you try to put around 10 disk drives on a scsi bus so this need to be looked into.

Good luck!