How to Reduce Impact of 'imsimta qm sum -data -oldest' on Job_Controller ?
(Doc ID 2454582.1)
Last updated on AUGUST 13, 2019
Applies to:Oracle Communications Messaging Server - Version 8.0.2 and later
Information in this document applies to any platform.
Received an alert email from msprobe:
It appears the above is due to a combination of two things:
- Started doing "imsimta qm sum -data -oldest" every minute so the age of messages in the queue could be monitored
- Started using filter_discard = 2 and system has accumulated over 20,000 messages in the filter discard channel
Wondering if '-data' (without '-oldest') makes a request which job_controller can answer from a summary data structure, but '-oldest' causes it to walk its entire in-memory queue structure to examine the dates. Is that correct ?
Why this is needed ?
We need to do "imsimta qm sum -data -oldest" because we need more fine-grain monitoring of our queues. Historically, we have only alerted when the sum of all the queues reaches a certain level. That can lead to false alarms. A large number of messages could be no problem at all if they are all relatively recent. A small number of messages could indicate a serious problem if they are all relatively old. Doing "imsimta qm sum -oldest" (without -data) might be a workaround, but that will create a significant IO load. So this is not a valid workaround.
How bad is the impact ?
The impact of "imsimta qm sum -data -oldest" taking several seconds and using a lot of CPU is not really too bad. Who cares if it uses a little CPU or takes a few seconds to get gather the required data ? However, if msprobe detects that job_controller is not responding during this, then doesn't that mean that processes doing enqueues are also having trouble communicating with job_controller? If that is the case, then the impact of this could be that messages get dropped into the queue without job_controller being notified. That would mean that our attempts to monitor the queue could be causing messages to sit in the queue for possibly a few hours before being tried for the first time. That could be a very significant problem if it happens to an important message or if it impacts customer communications.
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!
In this Document