We have a number of production servers running SUSE 9 Enterprise Linux. We have a custom software installation on our servers>
We need a simple system that will do the following:
1. Perform a function that will simulate activity.
Example: Send email to port 25 on Server1
2. Perform a test function that will result in good or bad
Example: Test for email sent to Server1 laning in POP3
mailbox on Server2.
If Failure then
a. log failure - server, servicetype, time - mysql log
b. perform sequence of actions to restart failing service
(Example: Perform openldap restart on server2, perform postfix
restart on server2).
c. perform simulation test function again.( like step 1 above)
d. check log for specific action.(like step 2 above)
e. failure (2nd time) = notify the world, call the police,
help me, I am dead.
3. Daily reports on any failures of more than two with the same server and servicetype.
The test objects can run on a centralized test server or on the individual servers. Or a combination of both.