Can't boot off newly created mirror

[Pete Randall]

I get "IODC ENTRY_INIT failed. status -3". I've gone through all the steps several times thinking I missed something (thought the other server I just did worked just fine). I'm beginning to wonder if the disk is bad.

Any thoughts?


Pete

[Steven E. Protter]

Hi Pete,

Why don't we try a different disk in that position? From one of the servers that isn't configured yet? Something like that.

Otherwise run the standard disk tests and see if you find something. xstm or cstm exercize comes to mind.

Next, I've seen a lot of instances where you get different firmware on to supposedly identical disks in the same machine. Bad quality control I guess.


You must have seen this, I've posted it often enough. Just putting it out there in case a brain molecule misfired.

pvcreate -B /dev/rdsk/c1t0d0 #use real disk

mkboot -l /dev/rdsk/c1t0d0
mkboot -a "hpux -lq (;0)/stand/vmunix" /dev/rdsk/c1t0d0 # use real disk


# mkboot -b /usr/sbin/diag/lif/updatediaglif -p ISL -p AUTO -p HPUX -p PAD -p LABEL /dev/rdsk/c?t?d?

If you are running 64-bit OS:

# mkboot -b /usr/sbin/diag/lif/updatediaglif2 -p ISL -p AUTO -p HPUX -p PAD -p LABEL /dev/rdsk/c?t?d?


vgextend /dev/vg00 /dev/dsk/c1t0d0 # same thing
lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/c1t0d0

# real disk. repeat for other lvols

lvlnboot -r /dev/vg00/lvol3 # root fs /
lvlnboot -s /dev/vg00/lvol2 #swap
lvlnboot -d /dev/vg00/lvol2 #swap/dump
lvlnboot -b /dev/vg00/lvol1
lvlnboot -R
lvlnboot -v
setboot
setboot -a 52.1.0 # second disk

Go get 'em

Regards,

Steven

[Michael Tully]

Pete, a man of your talents should know better ;^)

I would do as suggested by SEP, place another disk in the same spot and try it. You could even go to the extent of taking an make_recovery tape and using your live disk in the slot that doesn't work.

[James A. Donovan]

Does a search at the ISL prompt see the disk?

here's a couple links I found related to IODC

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=469390
http://ftp.parisc-linux.org/docs/arch/iodc.pdf

[A. Clay Stephenson]

Is this is a real HP disk with HP firmware on it? Use STM to compare the firmware level of this disk with that of your current boot disk. I assume that all the mirror steps were done and completed normally including pvcreate -B.

One other thought: Dumb pilot error. Does your boot path match the disk that you just mirrored?

Does SEA IPL list this device?

[Tim D Fulford]

hi

have you tried booting with "hpux -lm -lq" i.e. lvm maintanance mode & lost quorum?

I had a similar problm once but we had jiggled 'round with SCSI cables & it turned out disk was not terminated ...

Just my 0.02â ¬

Good luck...

[Pete Randall]

OK, that wasn't the enlightenment I was hoping to see this morning so here's the rest of the details - maybe that will help.

This is a brand new rp7410 with 4 36GB hot swap disks in it. The machine is ultimately going to be virtually partitioned into two. So, each of these virtual partitions will own 2 mirrored 36BG drives. I mirrored pair number 1 with no problem, then started on pair number two. After assuring myself on pair number one that I had all the steps right, I decided to be clever/lazy and use the following little script on pair number two:

export RAW=/dev/rdsk/c17t6d0
export COOKED=/dev/dsk/c17t6d0
export ORIG=/dev/rdsk/c14t6d0

pvcreate -f -B $RAW

vgextend /dev/vg00 $COOKED

mkboot $RAW
mkboot -a "hpux -lq" $RAW

mkboot -a "hpux -lq" $ORIG

for lvol in lvol1 lvol2 lvol3 lvol4 lvol5 lvol6 lvol7 lvol8 lvol9
do
lvextend -m 1 /dev/vg00/$lvol $COOKED
done


The script runs fine and all appears as it should.

I then went to test the mirrors by booting off them. Pair number one: no problem. Pair number two: "IODC ENTRY_INIT failed. status -3". This message, by the way, comes up long before you get to the ISL prompt so I can't do anything with maintenance mode or quorums or anything, Tim.

As far as swapping in another disk, I suppose I could take the mirror off pair number one and try that in pair number two. I guess I'll give that a go unless someone comes up with an inspiration this morning.

I find it hard to believe that a brand new server would have a bad disk in it, but I ran out of guesses on the third try yesterday afternoon.


Pete

[Robert-Jan Goossens]

Hi Pete,

Did you see this doc.

Document description: How to fix ENTRY_INIT boot failures
Document id: DFQKBCA00000112

Europe
http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000062687387

US
http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000062687387

Kind regards,
Robert-Jan

[RolandH]

Hi Pete,

seems to be a hardware problem.

Here is a document in the knowledge base, which describes ENTRY_INIT problems.

http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000063208946

Roland

[Jakes Louw]

Pete

Maybe I'm being dense, but where's your call to lvlnboot?

[Elmar P. Kolkman]

One other thing, we ran into too with something like this... Check if your root and stand are on the same physical extents on both disks. If not, it will not work, because this block numbers are used when booting the system!!!

To check this use lvdisplay -v, of course.

[Pete Randall]

The docs mentioned simply offer that it's an "error of an unknown type" - not much help.

There is no call to lvlnboot, because that's outdated - it's "unnecessary at version 10.x and beyond."

I'll check root/stand block numbers but I can't imagine how they could be different as long as things are mirrored in the proper order.


Pete

[Dietmar Konermann]

Pete,

the IODC's ENTRY_INIT points to the firmware's routine which initializes/opens the scsi bus and the boot device to read the LIF header from its beginning. Assuming that the LIF header and the LIF volume are correct and contain valid information, this is likely to be an hardware issue.

I found hints that this symptom could also be caused by messed up SCSI NVRAM contents. To correct this, you please try the command 'SCSI default' from the MP -> CO -> SERvice menue. (Attention, Initiator IDs get reset to 7 also!)

Best regards...
Dietmar.

[Pete Randall]

Dietmar,

Could you explain what that is supposed to mean? I barely understood the command you wanted me to run but I have no clue what "(Attention, Initiator IDs get reset to 7 also!)" means.


Pete

[Dietmar Konermann]

Pete,

no problem. :) I just wanted to point out that for SCSI shared bus configuratons (Serviceguard e.g.) you need to take care that you get no SCSI IDs conflicts with other nodes' initiaotrs. Typically the Initiator ID of one node needs to be set to something different than the default '7'. These setting would be changed back to default with the "SCSI default" command also.

Best regards...
Dietmar.

[Pete Randall]

Dietmar,

My ioscan shows pair number one as

0/0/0/3/0.5.0
0/0/0/3/0.6.0

and pair number two as

1/0/0/3/0.6.0
1/0/1/0/0/1/1.6.0

Would that give me a SCSI ID conflict in the second pair, despite the fact that they're on different LBAs?


Pete

[Dietmar Konermann]

Hi, Pete!

I'm just talking about the Initiators... the controllers ID. Check with:

# ioscan -fnkCctl

If this returns ID7 for all initiators then your are on the default and should get no conflict.

Best regards...
Dietmar.

[Pete Randall]

All 7s, Dietmar!


Pete

[Patrick Wallek]

Forgive me if this is a crazy suggestion Pete, but I know nothing about the rp7410's or how you partition them. The 2 disks you are now trying to boot from -- Are they assigned to the partition you are trying to boot?

It sounds as if you're just not seeing the disks which is the reason I ask.....

[Pete Randall]

Patrick,

Not crazy at all (at least not to me!).

The rp7410 is capable of hard, or physical, partitions if you have multiple cell boards, etc., which I do. However, I am not using physical partitions, all the resources are dedicated to one physical partition which I ultimately intend to split up with software, or virtual, partitions.

So, at this point, yes, the disks are assigned to the partition I'm trying to boot. Since I only have one physical partion and I haven't virtually partitioned yet, everything is.


Pete

[Pete Randall]

Dietmar,

Come back and get your bunny! Out of desparation, I tried your "'SCSI default' from the MP -> CO -> SERvice menu" command and it now works. I still have no idea what all that is about but at least I can stop tearing out what little hair I have left - for a while anyway.






Thanks to all for your help. The problem has been identified and the
solution has been found. Therefore this thread is:



_CCC______L__________OOO_______SSS______EEEEE_____DDDD
C___C_____L_________O____O_____S___S_____E__________D___D
C_________.L_________O____O_____S_________E__________D___D
C_________.L_________O____O______SSS______EEEE______D___D
C_________.L_________O____O_________S_____E__________D___D
C___C_____L_________O____O_____S___S_____E__________D___D
_CCC______LLLLL______OOO_______SSS______EEEEE_____DDDD




No further points will be awarded (unless, of course, I feel like it).


Pete

[Dietmar Konermann · Accepted solution]

Well, I don't like to come back without some added value. :)
http://www.dilbert.com/comics/dilbert/archive/images/dilbert22440040040308.gif

Have a nice weekend!
Dietmar.