As a PI administrator I would like to be able to create multiple instances of "PI Analysis Service" windows service for these reasons:
- avoid new potentially problematic analysis from affecting the ones that are operational and working correctly by setting one production instance and one test instance for example.
- every instance could have its own log files and we could move some analyses to a new instance to troubleshoot them
- make possible to distribute the resources usage for asset analytics on more than one server
The idea was already declined in "https://feedback.osisoft.com/forums/555148-pi-server/suggestions/33839161-multiple-instances-for-asset-analytics" nevertheless i think this should be considered for rolling out AF in big companies.
Same. Seems like an easy way to do this is to just have the calcs set a processing server. Then you could move things arround between multiple servers if you need to. I agree this is a big point of failure and does not allow for good scalabilty.
This topic was raised before: https://pisystem.feedback.aveva.com/ideas/PIDA-I-921 and had more then 220 votes, clearly very important for your customers and for our company. And still Aveva doesn't anticipate moving forward to enhance the PI Landscape in a full redandant, scalable setup for every component. Not even putting this important topic on your roadmap is very disappointing. Please reach out to your customers with a sustainable solution! Data Analytics becomes more and more important and therefor we have invested in a landscape that is high available built on modern infrastructure. The only component that is a single point of failure is the analytics-engine. So we would like to have a proper solution. *We would like to have a scalable environment for our Analytics Engine to properly manage larger systems like ours. - Multiple Analytic engines/processes running on the same time depending of the load of the calculations, events that needs to been taken care of; Seperate Analytic engine/process per AF DB (not AF-server!) *We also would like to have a high available setup of the Analytics engine where it can work with a redundant PI AF setup. - No outage during maintenance- windows - Automatic failover in case of unexpected outage of the Analytics Server The current setup with only one Analytics-engine is maintanable for small systems, but for larger systems it is a real problem. - A single user can roll-out too many eventbased calculation which kills the analytics engine - A quick restart of the calculation Engine results of backfilling all the calculations and could take up to +1hour ○ This restart is often required due to an issue in PI AF analysis The solution of setting up a seperate AF-server, in our case setting up an extra redundant pair AF-servers is not a sustainable path forward - More servers, is a higher operational cost ○ More maintenance, more effort, Server cost, licences,… - Management of global AF templates is more difficult - Splitting up a global AF hierarchy on multiple independant AF-servers is not desirable and feasible - Difficult for users to know which server to connect to
Multiple instances was supported with PI ACE but they seem reluctant to implement this as it is more complex to partition due to the integration with AF. While most of the PI system scales well Analysis/Notifications does not as you only have one engine so the need to follow best practices is important. I'm definitely in favor of multiple instance support including multiple instances on a MSFC.