Cannot create /var/adm/utmpx

风之幻想

楼主后续又做了那些操作呢?用cdrom进到单用户下.看能不能cp一个/var/adm/utmpx从光盘上到你的本地的启动分区上看看.

sundown

不能随便touch的,不然会造成无法登录,用光盘试吧

nononhm

SOLUTION SUMMARY:
With a Solaris software CD in the cdrom drive.
#1. boot cdrom -s
#2. mount /dev/dsk/c0t0d0s0 /a
#3. ls -l /a/devices/pseudo/sys*

      If there is NO MATCH - this is the key that the kernel patch
      install is the culprit. Proceed to step 4.

      If there IS a match - then it is likely that something else
      is responsible for the NO BOOT issue.  For example,
      /etc/vfstab was recently edited and the mount of /var is incorrect.

#4. Get the "postpatch" script from the 106541-05 directory.  It might
still be on the hard drive, or it can be moved on to the system via
ftp.  106541-07 uses the identical postpatch script and it
can be used as an alternative to the -05 postpatch script.  

#5.  #ROOTDIR=/a
   #export ROOTDIR
   #chmod 755 postpatch
   #./postpatch
  
#6.  ls -l /a/devices/pseudo/sys*
      should now see the devices:
sysmsg@0:msglog and sysmsg@0:sysmsg

#7.  #cd /
   #umount /a
   reboot

Note: To ftp the postpatch script from a server on the network, just
use the ip address of that server.  #ftp ip_address

If additional network configuration is needed to accomplish the ftp,
reference InfoDoc 19086 How to access the network from a "boot cdrom -s"
environment.  

The postpatch script is shown here for reference:
#! /bin/sh
#
# This is for the remote console.  It checks for existence
# of the new device nodes, and creates them if necessary.  This
# code emulates what would otherwise occur in SUNWcsd/postinstall,
# it is a requirement that this code run before the reconfiguration
# reboot.
#

# ensures an entry is in /etc/name_to_major
if [ -f $ROOTDIR/kernel/drv/sysmsg ]; then
      /usr/sbin/add_drv -n -b $ROOTDIR sysmsg
else
      exit 0
fi

devnode=$ROOTDIR/devices/pseudo/sysmsg@0:sysmsg

grep sysmsg $ROOTDIR/etc/name_to_major >/dev/null 2>&1
if [ "$?" = "0" ]; then
      major_sysmsg=`grep sysmsg \
              $ROOTDIR/etc/name_to_major | nawk '{ print $2 }'`
else
      exit -1
fi
rm -f $devnode
mknod $devnode c $major_sysmsg 0
chmod 0600 $devnode
chown root:sys $devnode

devnode=../devices/pseudo/sysmsg@0:sysmsg

devlink=sysmsg
( cd $ROOTDIR/dev ; \
      rm -f $devlink ; ln -s $devnode $devlink )

devnode=$ROOTDIR/devices/pseudo/sysmsg@0:msglog
grep sysmsg $ROOTDIR/etc/name_to_major >/dev/null 2>&1
if [ "$?" = "0" ]; then
      major_sysmsg=`grep sysmsg \
              $ROOTDIR/etc/name_to_major | nawk '{ print $2 }'`
else
      exit -1
fi
rm -f $devnode
mknod $devnode c $major_sysmsg 1
chmod 0600 $devnode
chown root:sys $devnode

devnode=../devices/pseudo/sysmsg@0:msglog

devlink=msglog
( cd $ROOTDIR/dev ; \
      rm -f $devlink ; ln -s $devnode $devlink )
exit 0

APPLIES TO: Operating Systems, Operating Systems/Solaris, Operating Systems/Solaris/Solaris 7, OS Install/Boot, OS Install/Boot/System Boot, AFO Vertical Team Docs, AFO Vertical Team Docs/Install/Boot, AFO Vertical Team Docs/Kernel
ATTACHMENTS:

nononhm

做完之后
ls -l /a/devices/pseudo/sys*
      should now see the devices:
sysmsg@0:msglog and sysmsg@0:sysmsg
这两个文件丢失了
/etc/name_to_major
/dev/console
这两个文件具体是有什么作用的？

sourit

原帖由 nononhm 于 2006-2-23 11:02 发表
恶化了。。。现在连单用户都进不到：
Rebooting with command: boot                                          
Boot device: disk:a  File and args:
SunOS Release 5.8 Version Generic_117350-11 64-bit
...

老大，我的也是这样，只是我是拔掉了做了RAID1的其中一块0#盘，从1#盘启动，就产生如此的状态，也可能是/var没有mount上来的缘故，是否在做了RAID1的情况下，从1#盘就不能启动了？

nononhm

单盘+系统，没任何东西。
回11楼，cp的时候很久都没响应，要ctrl+c才能操作

jobandmiss

这种问题的原因就多了，在出现这种现象的时候你都做了什么操做，什么机机型，硬盘是什么接口的？

hjp1973

Pls check /etc/vfstab ,if there is wrong or not?
Msybe ,you try to run "fsck -F ufs /dev/rdsk/c0t0d0s0"

[ 本帖最后由 hjp1973 于 2006-2-23 15:56 编辑 ]

nononhm

都做过了，不见起色。

lyh003473

出现这种问题。和var分区关系很大。在相关方面找吧