Hello guys! Can you suggest some software for doucmenting Servers?
thanks in advance
Yes - how about vi?
Seriously: without knowing any details it is hard to suggest anything.
Your question is like "i want to buy the best car - please advise": without knowing where you want to go with it (autobahn, offroad, racing track), how much cargo you have to transport (none, some trunks max, tons of it), where you want to go (in middle europe or the front line in Iraq), etc. this could be anything from a Ferrari, a Land Rover, an articulated lorry to a M-1 battle tank.
To paraphrase "Pill" Dr. McCoy: we are system administrators, not psychics. That makes us highly efficient information-processing devices. Put much in to get much out. Put little in and you get little out. Guess what you get when you put bullshit into this system.
I hope this helps (to enlighten you about our modus operandi).
bakunin
1 Like
.....................
With AIX there is prtconf in /usr/sbin it is a shell script and if there is info you want other than what is produces you could add or change it. But don't change the original. Make a copy and change as needed. The prtconf is used by the OS at time(s) and expects some info to feed other scripts.
Documenting what ? You can build a script which will capture some important data, which is then saved as a html formatted file and scp'd to your web directory.
One could use monitoring software like nagios or solarwinds. Nessus for security assessments etc.
Perhaps you could enlighten everyone here with an idea of what kind of information you are attempting to assimilate ?
Well what I mean is that I want some software where I describe my current environment describe what DB2 functions are used Describe TSM backup restore policy describe my current dual VIO configuration and so on.. the reason I want software like this is that If I leave my job I want somebody to understand how things are configured here through this documentation software.. because when I came here there was not any single explanation how things were working and I remember I really had difficult days.. I hope my explanation is OK now.. 
thanks everybody for your response I really find this forum interesting and cognitive as well
Now, as you have described what you want to do it is a lot easier to give advice. Have you said that before i would have written the following a week ago:
My suggestion is to set up a Wiki for this. You can group the articles via categories and this way create a structure, while retaining all the free-text capabiliies a database would deny you, With a database you will have to decide on the structure beforehand, with a Wiki you can (re-)design as you go. Start somewhere and over time the documentation will mature.
Still, some planning is not bad before you start to write: lay out a basic category structure and you might want to create some templates for more record-oriented data (servers, networks, SAN shares, ...). A server description template with properties like hostname, resources, OS version/release, IP address, etc. is quite similar to a table layout in a database: you define properties of a (class of) objects - here it is records in a table, there it is articles with an identical layout.
Setup of a Wiki is straightforward and takes typically some minutes, not more. An advantage is that the information is propagated via a web server, so you do not have to take it with you and the only tool you need to work with it is a web browser.
1 Like
I use both, mediawiki and self written html documentation,
but you may use dokuwiki, since it hast more security features, in mediawiki, everyone can read everything, but I don't want to migrate to dokuwiki, since we have 1000 articles+
here is my self written html documentation, wrote this a long time ago,
it's also very helpful to see how servers are configured, and how they looked like 2 months ago
#!/usr/bin/ksh
DATE=$(date +%y%m%d%H)
#choose OS from: solaris, aix, aixvio, linux
BINDIR="/usr/local/bin"
OS=aix
JOBLIST="${BINDIR}/joblist_$OS.txt"
HTMLDIR="/tmp/systemscan/$OS/"
HTMLFILE="$(hostname)_${DATE}.html"
HTML="${HTMLDIR}${HTMLFILE}"
echo $HTML
if [ ! -d $HTMLDIR ] ; then mkdir -p $HTMLDIR ; fi
>$HTML
echo "<html>
<head>
<title>Automated System report</title>
<style type=\"text/css\">
span {
display:none;
background-color : #ccffff;
}
body {
background-color : #ffffff;
}
</style>
</head>
<body>" >> $HTML
echo "<H3> Automated System Report for Host: $(hostname)</H3><br/>Date: $(date)<hr>">>$HTML
SID=0
cat $JOBLIST | while read line
do
SID=$(($SID+1))
DESC=$(echo $line | awk -F\& '{print $1}')
COMMAND=$(echo $line | awk -F\& '{print $2}')
echo "<hr><H4>$DESC</H4>" >> $HTML
echo "<a onmouseover=\"this.style.cursor='hand'\"
onclick=\"a_$SID.style.display='block'\">
<FONT SIZE=2>show</FONT></a>
<a onmouseover=\"this.style.cursor='hand'\"
onclick=\"a_$SID.style.display='none'\">
<FONT SIZE=2>hide</FONT></a><br/><hr>" >> $HTML
echo "<span id=\"a_$SID\"><hr><pre>$(eval $COMMAND)</pre><hr></span>" >> $HTML
done
echo "<br><H3> End Report for Host: $(hostname)</H3></body>" >> $HTML
then you need a joblist_aix.txt, may look like mine:
uname -a & uname -a
oslevel -s & oslevel -s
lparstat -i & lparstat -i
lspv & lspv
lspv -l all disks & lspv | while read i ; do lspv -l $i ; echo ; done
lsvg & lsvg
lsvg -l all vgs & lsvg | while read i ; do lsvg -l $i ; echo ; done
df -g & df -g
mount & mount
lsdev -C & lsdev -C
lscfg -vpl all adapter & lsdev -Ccadapter | while read a b ; do lscfg -vpl $a ; echo ; done
netstat -in & netstat -in
netstat -rn & netstat -rn
netstat open ports & netstat -an | grep -w LISTEN
lslpp -l & lslpp -l
lssrc -a & lssrc -a
errpt & errpt
errpt -a & errpt -a
/etc/passwd & cat /etc/passwd
/etc/group & cat /etc/group
/etc/hosts & cat /etc/hosts
/etc/security/user & cat /etc/security/user
/etc/filesystems & cat /etc/filesystems
left is the heading in the html file, right the command to be run
then copy the generated html file to a webserver
I run it every saturday on all hosts, and copy it to folders with timestamp
made a screenshot of the rendered html (attachment)
cheers
Edit: looks like the image is not attached
1 Like
In a Mediawiki it is possible to limit the right to view and/or edit to registered members. A customer of mine coupled this with an LDAP group and, honestly, i think this is all the security i suppose to be necessary for the documentation. Usually you don't put the documentation web server in a public network anyway so a username/password should suffice in most cases (and all cases i know of).
True, dokuwiki has indeed more security features, but - see above - i don't think they are necessary for the vast majority of cases. I have worked for banks, credit card companies and insurances and they didn't need this level of security for their documentation.
I hope this helps.
bakunin
from the media wiki site:
Extension:Page access restriction - MediaWiki
can you tell me which add on you are using bakunin? would like to check/use it.
cheers
I didn't say "per page restriction" - in fact it is an "all or nothing" type of restriction. The documentation i had in mind was for SysAdmins, DBAs and a few select people who had related jobs (some system architects for instance). They were put into a LDAP group "adminwiki" and this group membership was queried in the login (IIRC it was "login.php" - it was even hardcoded there). If one was member of this group he had full read/write access to all pages - just the wiki admin (i think in wiki syntax the "bureaucrat") with the right to delete pages, etc. was a separate account.
I will gladly provide the technical details, but it will take a few days - i am no longer in this project and have to contact the people there as well as ask their permission.
I hope this helps.
bakunin
I got you wrong, thought you mean per page restriction 
then I'm not going to need the information, but thanks for your offer!
our wiki is used from web developers as well as unix admins, dbas and so on,
so defining a per team area would be great for sensible data, even such as passwords