ServDoc-1.0rc1 is out.
Many internal changes, some code cleanup, some new modules (exim and postgres, more to come) and ServDoc is now easy to translate.
If there aren't any bugs, there won't be many changes to version 1.0
ServDoc is now multi-lingual.
For now, just English and German are supported.
I'm looking for volunteers to translate to other languages.
ServDoc Release 0.9 is out.
I think, ServDoc is now usable for most datacenters too. Many new modules and bugfixes since 0.8.
There is a table with tested configurations (please add your configuration) here.
We have working modules for MQSeries und UDB. (The same as with the oracle module: you don't have to configure anything. Just run Servdoc)
We have a bind module
ServDoc - What is it ?
The problem, part I
Imagine you have a production server with several software packages on it.
It is your job as sysadmin to document everything on that server so that you or your colleages are able to recover from every possible problem because you have all the information you need.
Oh, easy. You take your time (one day or two), you copy and paste some files and the output from some commands to a word document (or openoffice or whatever) and you are fine.
Ok, half a year later you do it again. And again and again.
And you do it for every server you have.
Depressing, isn't it ?
The problem, part II
No problem, you are an experienced sysadmin (you earn your money! and you are worth every cent) and therefor you are able to write a shellscript, that documents everything on that server. And you can run that shellscript from cron and mail that output to your account every week or so.
That is it. Till you install some new software on that server (and forget to adjust your shellscript).
ServDoc is the sum of all that documentation shellscripts on all servers in your data center, in our data center and hopefully in all other data centers.
Replace your shellscript with ServDoc and it will give you all the output you need (and maybe some more because you forget one or another thing to include in your shellscript).
You haven't to configure the tests in ServDoc. You even haven't to put the complete servdoc-directory tree to your server. You have just to copy the last version of the "self extracting and running" servdoc-Distribution to your server and run that. Say one (!) command line argument about your output format (ASCII, HTML or LaTeX) and that's it.
How does it work
ServDoc consists of a set of modules (and submodules and sub-submodules and so on), which will test the requirements and then output the needed information (or not: eg you haven't installed oracle on your server. Then you won't get some information from the oracle documentation module).
The ServDoc script itself just collects all the output from all the other modules and put that to one of the output modules.
The output modules take that information and bring it to a nice, hopefully good looking form.
At this time, there are three output modules: ASCII (simple ASCII-Text), HTML (a complete Webpage with Table Of Contents and so on) and LaTeX (with the possiblity to generate PDF).
Other output modules are planned (XML,maybe openoffice, maybe database import).
Can I see some demos ?The output format too is extensible. For now there are modules for ascii output and html output.
I would like to receive an email if you use ServDoc.
active work ongoing
History of ServDocMy first thoughts of a automated documentation system starts some years ago at a HP/UX troubleshooting training. Our first exercise there was to collect different information about our installation that we would need later on at the training like "at which blocks in your filesystem are your superblocks stored ?" or "How big are your filesystems ?".
At this training I wrote a short (Shell-)script especially for HP/UX.
Back at work I took that idea to more generality and wrote "printdoku" (the first version as a shell script, later on as perlscript). "printdoku" took a list of commands and headings and printed (on STDOUT) the heading and the result of that command. We had cronjobs to do that weekly and to mail that to one central "documentation server".
But - you had to configure that list of commands manually and that was difficult and time consuming and it was not easy to implement a new documentation item on all servers.
Therefor the idea of ServDoc was born.
What else ?
You can find a list of tools for this web site.
If you want to comment something, mail me.
Ulrich Herbst Last Changes: 23.02.2004 23:06:27