Hi All,

I’m not sure if this is the right section for this kind of thing so if it’s not, I apologise!

I would like to develop a RHEL server scan script which when run will scan the server and produce a output file with the results. So the scan would include stuff like:
disks, pv’s, vg’s, lv’s, cpu, ram, nic’s, nfs setup, packages, services, config setup, users, groups, etc
Basically a all in one snapshot of the server spec, after which all the data gets put into a neat file for the user to review.

I have written a bunch of individual scripts for each task over time but it’s really messy. At the time I wrote those scripts, I used whatever method was best to get it working. Some have function structures, others have loop structures and a whole variety of others in between.

Now I am looking to try and collate all those scripts into one nicely structured single script. What I’m hoping for is advice and tips from the community on how best I can achieve this.

The distro I’m currently working with is RHEL 7+ but I would like to make the script as universal as I can, so other versions of RHEL and maybe other distro’s like Solaris / HPUX.
The script it self is running bash shell.

Is this the right place to post my efforts and ask for help / improvements? Or does anyone know if someone has already written this kind of script I can use as an example?

Cheers

Take advantage of your experience so far to re-write each check as clean fn, with lots of error handling and have the 'main' section call each fn and concat the results into a file.

Ensure you use proper naming, indenting, comments in your code.

For extra points, you could get the main to run the fns in parallel (eg using '&" ) as they seem to be pretty independent, then it would have to concat the results together when all(!) have finished.

In either case, I'd definitely recommend using a dedicated dir to store the results in (maybe you want to store older versions of the results as well).

Also decide what format you want in the report eg csv.
To do that properly (harder than you think), consider using eg Perl which has modules for that & many other formats.

If you are also considering collecting performance stats, have a look at collectl https://www.tecmint.com/linux-perfor...collectl-tool/, http://collectl.sourceforge.net/index.html

Agree with chrism01 about Perl, specifically this module:
https://metacpan.org/pod/Spreadsheet::WriteExcel

If each of your scripts produces a CSV output (or CAN produce one), you can write the output of each script to a separate sheet in a spreadsheet. Makes it easy to look at later. And depending on what you want to do with the data, you can also shovel it into a MySQL table(s), and report on it from there. Check into Google Charts...free visualization plug-ins for web. Pretty flexible, although (as with most things Google), lacking in documentation for a good number of things. Simple PHP can query the database, and output a table, chart, graph, whatever.

Just a thought...don't know what your end goal is.

1 members found this post helpful.

I'm stricken with the similarity between the OPs specifications and logwatch.

1 members found this post helpful.

The request in #1 reminds me of the odd little utility/service that Tumbleweed uses to build an /etc/issue file at boot. (That and the "syscheck" utility on Tru64.)

None of the bits of information would necessarily be that tough to write. Depending on the distribution I wouldn't be surprised to learn that something already has been written that does this.