Solaris 2.6 /dev question

Hi all,

Thanks to all who responded

(Arthur Darren Dunham

Dan Stromberg

Roger Fujii

David Foster

James Cannon)

Roger Fujii sent me a CR/LF translation program, as well as the -850

option to send to unix2dos/dos2unix to avoid checking /dev/kbd.

The others all suggested that a cron job or softlink was changing

the /dev/null permissions. I did an ls -lR |grep null

and found no softlinks to /dev/null on one machine...

and the crontab for system accounts were all default (the ones

installed w/ os, patches) on a few of the machines having this problem

(including the one I did the ls -lR on).

So I have a workaround for one problem (/dev/kbd) but no real answer

for the /dev/null permissions. I'll have to keep an eye on that, and

I'll let everyone know if I find the answer.

Thanks,

aki

On Fri, 5 Nov 1999, Aki Sasaki wrote:

>Hi all,

>

>I searched the archives for this one but couldn't find an answer.

>

>I have several Solaris 2.6 machines, with the recommended patches.

>I have noticed that:

>

>1) /dev/null (/devices/pseudo/mm@0:null) changes permissions to

>crw------ 1 root bin 13, 2

>

>2) /dev/kbd (/devices/pseudo/conskbd@0:kbd) does the same,

>crw------ 1 root daemon 16, 0

>

>which breaks scripts... catting /dev/null or piping to it doesn't

>work for non-root. /dev/kbd gives warnings to unix2dos and

>dos2unix (cannot get keyboard type, us assumed) which fills up logs,

>and since the scripts are written to check the logs for any non-standard

>errors, it results in a lot of extraneous work and emails/pages.

>

>so i get called every few days to log in as root to the various

>machines and do a chmod a+rw /dev/null /dev/kbd.

>

>these machines are in the computer room, 8 machines to a video switch.

>i thought that might be the reason that /dev/kbd keeps disappearing

>but it doesn't explain /dev/null.

>

>has anyone else seen this?

>i'd like to avoid the huge kludge of having an hourly cron'ed chmod.

>

>please email and i will summarize to the list.

>

>thanks,

>aki

[2647 byte] By [CodeProf.com] at [2007-12-25 11:24:00]