By John Alvord, IBM Corporation
I was working with a customer which had a new processing environment of Windows servers being brought up. They had defined two new remote TEMS and there were 450+ Windows OS Agents reporting. The reported problem was that too few events were arriving at the operations center. After considerable work we determined that the remote TEMS was operating normally. There was some problem with the new processing environment such as a mis-configured network attached storage device which was preventing tasks from running normally.
The operations staff had a hard rule for the low event arrival and that was to phone on-call support staff. Operations was doing that every single night and the customer had not had a full night’s sleep in some days.
Situation to provide a full night of sleep
Since we were convinced the remote TEMS was not at fault, I created a situation IBM_15min_flipper which would create and event and close it 4 times an hour. The example situation is here.
The formula uses the LocalTime attribute group:
Every minute the formula is checked. From minute 00 to 07 it is true. From 08 to 14 it is false and so on. One interesting note: a test with Minutes <8 did not work because that field is a 2 character text field and not a number [as I imagined].
The situation was distributed to two Solaris Unix OS Agents. That provided 8 events per hour which prevented the operations staff from worrying.
The situation was a grand success. The customer contact had a good nights sleep with no calls from the operations Center. In the meantime the new processing environment is being worked on.
Sitworld: Table of Contents
Note: My wife and I in the Big Sur River, behind the River Inn Restaurant.
They set a few chairs and tables out in the river during the summer when the water level is low.