Has anyone been able to successfully implement process monitoring through Optimize on an Clustered IS enviornment with Load balancing?
The behavior I’m seeing is that steps is my process are occurring in a different sequence in Production. Throughout development and testing, the only difference has been the lack of clustering and load balancing. I suspect that the clustering between 2 IS’s may be delaying the completion of certain steps.
I have not used clustered Integration Servers but certainly have two Integration Servers logging against the same database. Though, theoritically I expect the same behavior with Integration Servers, I have not experienced the problem that you have outlined with the 6.1.5 implementations on windows/Solaris/AIX platforms. I have seen the behavior with Integration Server 6.5 even more vulnerable and I have an open case with TS right now on a particular model/monitor issue for 6.5
I have a reproduceable problem that I am working with TS to resolve.
Can you provide more specifics on your implementation including the versions and operating systems, etc?
We have figured out our issue.
What had happened was that both of our IS’s internal timestamps were out of sync and over the course of a week moving further and further out of sync until a weekend server recycle.
In order to resolve the issue, we are pulling timestamp off of a database which is synched to a timeserver.