When i try to boot my solaris 2.6 E220R server, i get this error message
> INIT: cannot create /var/adm/utmp or /var/adm/utmpx
>INIT: failed write of utmpx entry:" "
>INIT: failed write of utmpx entry:" "
INIT Single User Mode
Press Ctrl-d for normal boot process or enter roots password for system maintenance.
Now when i do either of the last sentence above, nothing happens, it appears to hang.
I used the solaris cd to boot to single user mode using
boot cdrom -s
with the intention of running fsck on the /var files system where the utmpx file recides. But when it boots to root prompt-single user mode with cd , i am unable to type successfully. Any letter i type just keeps running on , repeating it self like;
fffffffffffffffffffffffffffffffffffffffffffffffssssssssssssssssssssssssscccccccccccccccckkkkkkkkkkkkkkkkkkkkkkkkkk........etc.
Please how do i get out of this mess? I really need a quick response, it's my mail system.
As far as your booting problem, it's a documented problem -
Installing Solaris 2.5.1 11/97 or Solaris 2.6 5/98 on Ultra 5, 10, 60, 80 or E220r and E420r
with 440 or 450 Mhz CPUs, whether interactively using a CDROM or doing a
jumpstart install will fail with the message "hme0 - link down".
[b]Another failure symptom is multiple keyboard input repeats, or repeating keystrokes[\B]
.
To prevent this, the Operating Environment Installation CD p/n 704-6914-10 (Sept 99) or
704-7076-10 (Feb 2000) must be used. Complete procedures are documented in
the manual(s) 804-6657-11(Sept 99) or 806-4005-10 (Feb 2000)
"Installing Solaris 2.5.1 or Solaris 2.6 Software on 419+ MHz Systems" or
"Installing Solaris Software for Selected Hardware"
Look for the OECD and documentation in the shrink wrapped packet
that says "Binary Code License" or inside front cover of
"Installing Solaris Software for Selected Hardware" manual.
As far as your /var/adm problem, once you do get to single user mode (from cdrom - since it is complaining that it cannot create or write to the file, it means you will never log in by booting from the disk), check the permissions on the directory and on the file, if it exist.
$ ls -ld /var/adm
drwxrwxr-x 6 root sys 512 Aug 11 03:10 /var/adm
$ ls -l /var/adm/u*
-rw-r--r-- 1 root bin 828 Aug 11 19:57 /var/adm/utmp
-rw-r--r-- 1 root bin 8556 Aug 11
19:57 /var/adm/utmpx
(Reminder - you will be mounting your var partition on /mnt or /a - change your command accordingly)
Sorry, i made a mistake, its solaris 7 not 2.6.
I did boot to single user mode with the cd and i did run fsck, fixed errors, mounted the /var partition on /a, rm the utmpx, utmp, wtmpx, wtmp files, touched them and gave the default group and owner permissions. Then i restarted with init 6 ( after umount /a) but when it was booting i noticed an error message; "cannot sync file systems.....give up.....rebooting" and then i end up in square one.
Please what could be the problem? ....i think i have tried everything.
$ ls -l /var/adm/u*
-rw-r--r-- 1 root bin 828 Aug 11 19:57 /var/adm/utmp
-rw-r--r-- 1 root bin 8556 Aug 11
19:57 /var/adm/utmpx
On my system, after "bin" i have 0 for both utmpx and utmp ( not 8556 and 828) . What is the implication of these numbers? and what effect could it have on file performance?
Since this is a mailserver, do you have /var/spool/mqueue as a separate partition? If not, it could be that your /var partition is full and that is why you can't write to it.
Take a look at the link above. But that apart, i do have large /var/spool/mailbox directory but when i do df -k, i still appear to have quite a lot of space . BUt i am cleaning out my mailbox directory nevertheless.
Could this problem really affect the performance of this server in this manner, and if soo why?
This was an important piece of information (Solaris 7 Jumbo Kernel Patch). Did you apply it and [I}then{/I} you could not boot? Did you apply the patch in single user mode?
Please answer these questions.
Also, please post the output from df -k.
If there is still space on /var/adm (whatever that partition is) then it should still write to the files. Another possiblity is that you have run out of inodes. Post your output from df -k.
no i did not apply the patch, i am not sure i know how to go about it, seems rather complicated.
I was trying to see if i still had the patch, locally on the disk soo i wont have to copy to diskette ( E220R dont even have diskette drives) etc, but after mounting /, i couldnt seem to find it.
I cannot copy the output of the df -k for u as the machine is off the network since it cant boot, but this is a transfer of the handwritten output below;
# mount /dev/dsk/c0t0d0s0 /a #cd / #cd /a #df -k
Filesystem : /dev/dsk/c0t0d0s3
Kbytes : 12100380
used : 876196
avail : 11103181
capacity : 8%
mounted on : /a
Offcourse, on the server, this entry, along with others, is displayed horizontally.
With reference to chttp://sunsolve.sun.com/pub-cgi/retrieve.pl?doc=fsrdb%2F20768&zone_32=cannot%20write%20to%20utmp%20and%20utmpx%20files%20solaris%207 ,
i could not seem to find any matches when i did
ls -l /a/devices/pseudo/sys*
and i cant seem to find the post patch script 106541-05 either.
Please how do i go about this? i might be missing something.
Is there a chat room you belong to? i would like to chat with you as i try to resolve this problem, for advice. I am :: email removed :: and :: email #2 removed ::
Sorry, no, I don't use chat rooms - and unfortunately, I was not on this site this morning so I did not see your last message until now (can't use chat at work).
From what you posted - you didn't apply the patch.
This means the problem occurred due to the patch (according to the documentation). You must do Step 4 and below in the documentation. I've never used ftp to get a file onto a server booted from cdrom, but the documentation makes it sound like it is possible. Put the patch onto a disk (mount /dev/rdsk/c0t0d0s0 back to /a). ftp from your 'down' server to one that has the patch. Get the patch and then follow the rest of the steps.
