Really simple script needed to so the following.

We run both CentOS and Ubuntu here and I need to put together a single shell script with some simple logic to tell the OS's apart and run the appropriate patching/updating commands once that's done and then a reboot. The whole thing will get fired off by crontab at a given date and time for each system.

I've got a few ideas to tell each apart:
For CentOS I found this:

more /etc/system-release
CentOS release 6.9 (Final)

and for Ubuntu I found
uname -a
Linux cnxatlsgi01 4.4.0-75-generic #96-Ubuntu SMP Thu Apr 20 09:56:33 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux


And the update script without the logic is:

#!/bin/bash
#
#
#
#CentOS
yum -y update
#
#
#
#Ubuntu
apt-get update
sleep 3
apt-get -y dist-upgrade
sleep 3
apt-get -y autoremove
#
#
#
reboot

but I'm not sure of the best way to do the OS selection or if there is totally better way to go about this. We can't do any automated updating due to processes running the background (fuse) that really hate stuff messing with them when up.

Any help as always appreciated.

George

In Ubuntu, you have a short utility to get the release information:

That can be processed with sed or awk to extract what you need to make the decision.

See also about using command substituion in your script.

2 members found this post helpful.

uname -a

Why?
On the centos server, set up your cron to execute yum

On the Ubuntu, set up your cron to execute apt-get

AFAIK, a reboot is not necessary following an update on CentOS...but I could be wrong about that.

1 members found this post helpful.

i also think the approach is wrong.

but you have to look at /etc/*release* in each distro.
also packages like lsb_release might exist for other distros.
also a web search...

2 members found this post helpful.

As post #5, @ondoho

1 members found this post helpful.

Normally you have to reboot if a new kernel had been installed for that new kernel to be the active one.

WE have A mix of different types of servers and want one script we can push from a central source to all of then to run at the same time via a cron command where the script will decide which update command to run....now the dirty way to do it just to let it fall through and fail on the systems that don't have the other OS update utilities but I'm trying to not do that. We need and require reboots after every update run here in case a kernel is updated.

I must have been half asleep when I wrote the original request, or the notebooks screen format confused me. The utilities are fine for OS selection:

more /etc/system-release
CentOS release 6.9 (Final)

uname -a
Linux cnxatlsgi01 4.4.0-75-generic #96-Ubuntu SMP Thu Apr 20 09:56:33 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux

What I need help with since I know no sed/awk is the selection and if/then portion to decide which group of commands to run based on the given OS type and then end with a reboot and if neither if found exit with out a reboot.

if you want to know if the current os is either centos or ubuntu you need only to grep centos in /etc/some-file (or similar).

Or simply check for some string in some file.

That's very cute. Yes, I know that and showed it in the post above. What I needed help with is the logic for the if/then and awk etc.

That's very cute. Yes, I know that and showed it in the post above. What I needed help with is the logic for the if/then and awk etc.

https://linuxacademy.com/blog/linux/...if-statements/ ?
http://bencane.com/2014/01/27/8-exam...t-you-started/

1 members found this post helpful.

Pan Thank you! looked all over for this!!

I think the first will do what I need to get this done, need to look at it later, the second does not resolve, 404 error - sorry, looks like its incomplete....ahhh...you went and fixed while I was typing thank you second one there is exactly what I need.