I have an a homogenous environment of about 200+ server (supporting many accounts) with different application running on them.
and I also have to work on boring , mundane Incidents ... which are mostly related to configuring application and troubleshooting and reporting/qureying data.
While solving these incidents most of the time gets wasted for waiting the application console to load on a central server (for configuration tasks ...that can be automated) and taking multiple (4+) RDP/ssh to reach servers (most just to check/troubleshoot connectivity)etc.
All this can be done with batch or ssh script but they are very difficult to manage. And I every time i write a new solution to problem , it has to be pushed to these 200+ server, again a cumbersome task..And it becomes more difficult when you have complex scripting logic for adhoc query/task
I think puppet or Mcollective is right tool for me.. But i am not sure Which one is gonna help here.
I aim to design a system the where if an alert (on single or multiple hosts) occurs , it gets automatically resolved...or it could fetch diagnostic reports from the server or if it requires manual invervention for sysadmin stuff, i could do it from central console
I m guessing that if need network stats/ metric from all the host, MCollective can help but What i am not sure is whether Mcollective allows me to target each and every host individually and then run tasks on it.
↧