Can not boot

[Steven Lam_2]

Hi,
I just got the HP 9000 series that whenever it boots, it just got the errror messages as:
Can't find /stand/vunix or /vunix. And it comes to the IPL> prompt. Any hints or suggestions are greatly appreciated.

Steve

[Patrick Wessel]

just a few minutes ago:
http://forums.itrc.hp.com:80/cm/QuestionAnswer/1,1150,0x7809a24d9abcd4118fef0090279cd0f9,00.html

[Kofi ARTHIABAH]

Steven, it would help if you provided more info ... what kind of 9000 is it. Is it possible that before you got it, it had been decommissioned - ie. OS and all software removed? you can try doing a

IPL> sea

to search for bootable devices and then booting from one of the ones that show up.

[Madhu Sudhan_1]

It looks to me like the boot string is not configured properly.


At IPL type
1) sea
2) boot
3) Interact with ISL ? Say Y
4) ISL> hpux -is (;0)/stand/vmunix

If everything goes fine at this point and you boot to single user mode, then use mkboot command to overwrite boot string whcih set as vunix instead of vmunix.

Hope this helps.
...Madhu

[Anthony deRito]

Your at the ISL prompt. This means that you have performed all your hardware self tests, read stable storage and loaded the ISL from the LIF area. The ISL was then executed and defines, loads and passes control to one of several possible secondary loaders. The one your interested in is the hpux secondary loader.

From here you want to use hpux to list the available kernel files in stand.

ISL > hpux ll

Check to see if you have a vmunix or a backup. If the file is there, it may be corrupt. If it not there then you have to find a backup. If you do not have a backup file then you need to go to the CD.

If the comand hpux ll returns an error, the most likely causes are:

1) The LVM pointers to the root volume are corrupt - can be fixed using recovery CD and and selecting the option to rebuild only the bootlif.

2) The boot filesystem is corrupt - can be fixed by reinstalling the operating system.

If hpux ll succeeds then you know at this point lvol1 is OK on your boot disk. Now you need to check if your boot string looks OK. Run the command:

ISL > hpux show autofile

You will probably see something like:

hpux disk(8/4.5.0;0)/stand/vmunix

If you have a backup kernel, run exactly what is in the autofile except replace vmunix with the name of the backup kernel.

I'll stop there for now. Let us know how you are progressing.

Tony

[Steven Lam_2]

All,
Thanks a lot for your support. It still can't boot. I am sorry for the typo. I only got the ISL> prompt not IPL> prompt. My server is HP 9000 series J210. The OS is 10.20. I just got the server from the stock and need to fix it and I don't know what happened to it before. Here what I did:
1. At ISL> search -> errror (can't search)
2. At ISL>hpux -is or hpux -lm -> error (can't find vmunix) and came back to ISL> prompt.
3. At ISL>hpux -is (;0)/stand/vmunix or hpux ll -> can't work either.
4. I boot from CD and go to the Shell. At the shell I can't use any command except "cd" command.

I got another HP UNIX with the same model and OS version. I copy and tar the vmunix to the tape but I could not copy from the tape to the server. At the shell I got the root prompt but could not do anything. Thanks a lot for the support.

Steve

[newby]

Steven,

after trying to boot interactively, boot from the alternate path and stuff, which it looks like you've tried, I'd suggest that you will save time by obtaining a 10.20 Boot cd from HP. You should get one under your service contract and I'd reconfigure this host from scratch, It will be faster in the long run than to do any more troubleshooting.
- just a suggestion.

[Anthony deRito]

Steven, you cannot boot into single user mode from your boot disk if your kernel is not present or if it corrupt. Also, if the LIF volume on your boot disk is corrupt, then you would also not be able to use the secondary loader (hpux).

The way I see it is you still need to determine 2 things:

1. Is your LIF area corrupt on your boot disk.
2. Do you have a bootable kernel.

I wanted you to rule out #2 by executing the command ISL > hpux ll. You say it does not work but you did not indicate what the error was. This can take time to run so you must be patient. This is the only way to check to see if what you have on lvol1. lvol1 is where /stand resides and thus the kernel. This will also verify to a degree if lvol1 is OK.

Keep in mind that the two problems (a bad LIF volume and/or bad or missing kernel) both have to do with two different areas of your boot disk. If you draw a picture of your boot disk, it would look like this:

--- LIF header ---
--- PVRA ----
--- BDRA ----
--- LIF area (ISL, AUTO, hpux) ----
---- VGRA ----
--- lvol1 (stand) ----
--- lvol2 (swap) ----
--- lvol3 (root) ----

Also each problem requires a different approach to solve the problem. If you want to fix lvol1 assuming it is somehow corrupted, you need to boot into the recovery shell like you did by selecting option (x) from the main menu. (However there are a lot more steps you need to do than you did) You can then go ahead and get a kernel back but remember, this is a mute point if you have a bad LIF volume. If you want to onlt restore a bad LIF volume you would select option (r) from the main menu and then option (c) Rebuild only the boot LIF. So what is one to do???

I would chose to rebuild the boot LIF first if you are not sure. Then, when you boot into a recovery shell, all you want to do is run the command:

#chroot_lvmdisk

This command will create /ROOT, fsck lvol3 and lvol1 on your boot disk and mounts them to the new /ROOT.

Let me know how it goes and we can take it further. I know many people do not respond to comments so I will wait for your signal.

Tony

[Steven Lam_2]

Hi all,
I just got the system back by using recovering the bootlif from the CD Rom. However, I still have the problem with SAM under the kernel parameter. It shows a long error as:
Internal error: the kernel discovery tool used by SAM fail. Kernel changes cannot be made by SAM... The last line of the error is:
getkinfo: Unable to read the kernel file "/stand/vmunix". However, I can boot up the machine and login without any problem. I think the kernel is unstable. I could not run the command: /usr/sam/lbi/getkinfo -b.
Here are some other commands that I tried to run:
ISL>hpux ll -> It showed me all of the directories and files at the root (/).
ISL>hpux show autofile -> got three lines below:
Show autofile
:disc(8/0.6.0;0x800000)
AUTO file contains (hpux)
I don't know how to fix the kernel parameter problem.
Thanks a lot for your supportting.

Steve

[James R. Ferguson]

Steven:

You don't give the full text of the syslog message. Missing drivers and unsatistifed symbols will yield these types of errors when 'getkinfo' is run. For some analysis and repair guidelines you might want to look at documents #TBRIXKBRC00003312, #KBAN00000153 and #KBAN00000216.

...JRF...

[Matthias Zander]

Try to make a Kernel with mk_kernel:
mk_kernel -s /stand/system -o /stand/vmunix.test.

If it works then:
cp vmunix vmunix.prev
cp system system.prev
mv vmunix.test vmunix
reboot

[Anthony deRito]

Steven, now you know how to retore the boot LIF volume! Keep this in your bag of tricks... it saves a lot of time if you need to just get a system up and running as fast as you can.

Actually when you ran the command:

ISL>hpux ll

It does not show you all of the directories and files at the root (/) filesystem but the directories and files in lvol1 (/stand). You should have seen your kernel and/or backup if you had one. As I said before, this is a great way of checking if lvol1 is OK to the extent that your files are there. Now the fact that you had bad LIF volume tells me that there is some form of corruption on your boot disk. This may be where the errors are comming from. I would not trust your boot disk but thats your choice.

Tony