Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations strongm on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Boot Problem with 220R Entreprise from Sun MicroSystem 2

Status
Not open for further replies.

bduclos

IS-IT--Management
Dec 6, 2004
7
FR
Hi,

I have a boot problem with 220R Entreprise from Sun MicroSystem.
When I boot I have a message like this :

***********************************************************
boot device: disk file and args:

WARNING: forceload of misc/md_trans failed
WARNING: forceload of misc/md_raid failed
WARNING: forceload of misc/md_hotspares failed

INIT: cannot create /var/adm/utmpx
INIT: failed write to utmpx entry: " " (*2)

INIT: single user mode

type CONTROLE-D to proceed with normal startup, (or give root password for system maintenance)
************************************************************

When I try to write the root password or press CONTROL-D, I have nothing, like the keyboard is not connected. The only action I can make is to press STOP-A. The commands I know are devalias and printenv. I try to boot via the net but not answer.
In result I connect this machine to a terminal on the serial port A at 9600 bauds, 8 data bits, 1 stop bit and never parity. I send messages when I reboot the server but always nothing on the terminal console.
Perhaps, It's cause by a hot keyboard switch off. I don't know exactly.

Does anybody can already have the same sum and help me to diagnose and resolve my problem ?

Thanks in advance for your solutions. Benjamin
 
The warnings are from your meta devices you have set-up.

Not sure why you are getting a /var problem.

For the serial connection, are you connecting to serial port A through a Null Modem?
 
don't care about the warnings, it's a "bug", for SDS it's a warning if you do not setup hotspares, Raid5 or transdevices; if you like you can #-out these forceloads in /etc/system

I guess you are having problems with /var filesystem

You have to fix the Keyboard Problem first: disconnect keyboard and monitor if you use the Terminal on serial port!
If you use keyboard/mouse/monitor powerdown machine (keyswitch), hit STOP-N on Sun-Keyboard and power up the machine (best practise: ask a colleague for assistance, since you dont't have enough hands to do this), OBP writes something like "L1-N pressed, resetting to factory defaults"; next OBP tries to boot OS/Solaris; if keyboard works fine enter root-password to run fsck on / and /var

if this does not fix the problems post again, we can disable SDS, if this causes the problems...

You can boot from CDrom, mount the bootdisk and take a look at /a/var/adm/messages, maybe you get a hint!

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years
 
Like you advise me, I try to fix the keyboard problem without results.

1) The terminal doesn't respond with keyboard and screen disconnects.I have not any signals.
2) I try to press STOP-N during the start process but I only get this message :

Resetting NVRAM defaults because of L1-N keyboard command
Setting NVRAM parameters to defaults values.

and the keyboard does not respond too.

- I'll be interesting with your manipulation in order to disable SDS.
- The process of the CDROM mounting will be help me too.
Perhaps the main command is "setenv boot-device= CDROM"

I'm sorry but like you certainly see I'm novice in Sun Server Administration

Thank you again for your services. Benjamin
 
Benjamin, to boot from cdrom, use this faq as a guide:

Faq60-1568

This should be enough for you to examine any messages as suggested above. Post back with these if it's not clear what's wrong.
 
I have successfully boot from CDROM. That OK for that :)

1) Before mount action I have control with fsck like this : fsck /dev/dsk/c0t0d0s0 but no errors are detected.

2) When I mount the root filesystem with "mount /dev/dsk/c0t0d0s0 /a" I have my arborescence but I have nothing (not any hidden file too) in /a/var.

At this point of my crash system, I guess my problem is/are filesystem error(s). For this reason, I ask this specifics questions :

- How can I diagnose my issue ?
- How can I execute a completeness fsck on my primary boot disk ?
- How can I start my Solaris 5.8 OS (without reinstall system) by booting on CDROM device ? Is there any rescue mode to catch a problem ?

Thanks, for all your proprositions.
 
I would think that /var is just a mount point for the var disk partition. You can look at /etc/vfstab to see how your file system is setup.
 
just like bfitzmai points out, look at /a/etc/vfstab, was /var another slice/filesystem? Mount this...

- How can I execute a completeness fsck on my primary boot disk ?

you have to run fsck on every single filesystem, depending on how compley you installed it this is more than one command/step...

- How can I start my Solaris 5.8 OS (without reinstall system) by booting on CDROM device ? Is there any rescue mode to catch a problem ?

booting from CDROM is more or less the rescue mode, you can do almost everithing to heal the OS on Disk; to give further advice we need to know what is the problem (-> messages)


Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years
 
In fact when I'm at root arborescence / and I execute command "cd var" I'm next in /tmp/root/var and in /tmp/root/var/adm there is nothing. I have edit my vfstab file and it looks like this :

# device device mount fs fsck mount mount
#to monut to fsck point type pass at boot options
#
#/dev/dsk/c1d0s2 /dev/rdsk/c1d0s2 /usr ufs 1 yes -
/proc - /proc proc - no -
fd - /dev/fd fd - no -
swap - /tmp tmpfs - yes -

swap - /tmp tmpfs - no -
/proc - /proc proc - no -
/devices/pci@1f 4000/scsi@3/sd@6,0:b - / ufs - no ro
fd - /dev/fd fd - no -

- Can someone explain me what this file means ?
- Is there any problem in this config filesystem ?

Thanks for all.
 
this is the /etc/vfstab from the CD, your vfstab is in /a/etc/vfstab

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years
 
Had a similar problem on a v480. The root file system became read only despite saying it was read/write. Only way round this was to install from cd to single user to ufsdump the root system files as a back up. We then had to undertake a full system install from cd making sure to make /var and /usr as seperate slices to root. Good Luck
 
I have seen similar errors with this some came from the /usr partition not being mounted. Also if you are using SDS verify your /etc/vfstab like several other people mentioned. The last thing is SDS could be the problem I have had instances where I had to disable it then reinitalize it again, it is pretty simple process. You comment out SDS stuff in /etc/system, change your /etc/vfstab back to the original, and under /etc/lvm there is a file that contains the metadb config. You should then be able to boot of the regular root disk and redo your SDS config.
 
My file system setup indicates many informations but when I try to execute fsck on a device It says me that :

- can't stat /dev/dsk/c0t0d0s0

Before mounting the root partition I check my disk with fsck
and there is no problem.
I'm completely lost. I'm not able to restart the server because off the lock keyboard at the boot menu. It's very strange. Does anybody have already the same problem ?
It's probably nothing. I remember that I had change my shell before rebooting csh --> bash. Does this feature has an incidence with my problem ??

Thanks at all
 
default shell (root) should be /sbin/sh since it is
Code:
# file /sbin/sh
/sbin/sh:       ELF 32-bit MSB executable SPARC Version 1, [b]statically linked[/b], stripped
#

fsck: alwas check the /dev/rdsk/.... Device

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years
 
Yes That's OK. :)

I have booting Solaris 5.8 successfully.
The problem was bring about the modification of the shell.
In /sbin I had renamed sh in old_sh (happily) and rename a bash shell in sh. I guess that the bash shell shouldn't be mount at boot. I'm reassure that it is not a critical hardware problem.

I dismiss everybody who has purpose me a solution.

Thanks. Benjamin
 
depending on the way you partition your boot disk the libraries for the shell may be on a (not mountable) slice; you may change the default loginshell to any shell you like, as long es it is statically linked and available from the / Filesystem (not a separate eg /sbin FS)

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top