Requirements:
- Recently a client had a request to have the possibility to set a server/device in maintenance via a simple "nimalarm" message. (he used to do the same in NSM with a cawto command)
- A second requirement was that in the message he would set the servername and the # of seconds that the device needed to stay in maintenance.
The agreed alarm format was: nimalarm -l 1 "maint_on servername #seconds"
Solution:
The solution exist in a new custom probe, that need to be installed only on the main UIM hub: probe_maint_on.
A Nas AO profile will trap all alarms starting with "maint_on" an repost them with subject: "maint"
This probe will connect to a user defined attach queue: maint.
For each message we will create a new maintenance calendar and define the server into this calendar. When the definitions ends ok the original maint_on alarm is cleared.
The attached zip contains:
- probe_maint_on.zip version 2.5
- doc file
- sql examples how to delete empty and expired maintenance calendars
What's new in 2.9:
- you can now remove a server from maintenance by setting the #seconds to 0
- several extra checks to avoid creating empty calendars
- extra checks if sql login is ok
- if a server has multiple cs_id's, we will use them all
- probe is now installed in /probes/custom/probe_maint_on
(the attached zip 2.9 contains also a file: probe_maint_2.9.zip; this is the package to be added in your archive)
Tested with UIM version 9.0.2 on Windows
What's new in 2.91:
- tested with UIM 9.1
- reviewed the doc and install archive package to be as simple/clear as possible. The zip file posted is also the zip file to place in uim archive
- added some extra parameter tests