One of the problem I used to have as build & release engineer is to manage login to huge number of boxes through my Linux system. At the scale of 5-10 machine it’s a not a big problem but once you have close to 100+ boxes then it is not humanly possible to remember the ip’s of those boxes.
The usual approach for this problem is to maintain a reference file, from where you map machine name with the ip & find the ip of the box from this file, but again after some time this solution seems to be not that efficient. Another solution is to have a DNS server where you can store such mappings & then you can access these machines using their names only, this is the idle solution but what if you don’t have DNS server also still you have to execute the ssh command ‘ssh user@machine”.
I developed a simple solution for this problem, I created a utility script connect.sh, this script takes machine name as an argument & then we have multiple conditions statements which checks which ssh command to be executed for the machine name.
#!/bin/bash
if [ “mc1” == $1 ]; then
ssh user@
elif [ “mc2” == $1 ]; then
ssh user@
elif [ “mc3” == $1 ]; then
ssh user@
.
.
.
fi
This solution worked really well for me as now I’m saved from executing whole ssh command, also for machine name I’ve followed a convention i.e _ for example the entry for a machine for release environment that is hosting an application catalog the machine name would be release_catalog, similarly dev_catalog, staging_catalog, pt_catalog.. so you don’t have to remember machine names as well :).
