Forum Discussion
On premises data gateway upgrade (cluster)
- 5 years ago
Hey gaupop13 ,
disabling a member of a gateway cluster is simple, use the gateway dialog to stop the service, then update the cluster member start the gateway member - done.
You have to be aware, that some things changed over the last two years, make sure that your gateway machines meet the latest requirements. There have been some changes in the past: https://docs.microsoft.com/en-us/data-integration/gateway/service-gateway-install
Depending on the number of data sources that you are supporting, maybe you want to consider creating the new cluster first and then move the existing data sources to the new cluster. Maybe there will be a downtime for a certain dataset, but this can be reduced to a minimum.
Make sure you have all the required information and then simply create a new data source name:
- Name of the server
- Name of the database
- Service-Account and of course the password
- Users of the data source
I would recommend this approach, and prepare the workspace owners to change the connection accordingly using the new data source name, this will reduce the downtime (the unavailability) for directquery / live connection or dataset refreshes to a minimum, if this switch will happen between two data refreshes or during an off-time.
Hopefully, this provides some ideas on how to tackle your challenge.
Regards,
Tom
Hey gaupop13 ,
disabling a member of a gateway cluster is simple, use the gateway dialog to stop the service, then update the cluster member start the gateway member - done.
You have to be aware, that some things changed over the last two years, make sure that your gateway machines meet the latest requirements. There have been some changes in the past: https://docs.microsoft.com/en-us/data-integration/gateway/service-gateway-install
Depending on the number of data sources that you are supporting, maybe you want to consider creating the new cluster first and then move the existing data sources to the new cluster. Maybe there will be a downtime for a certain dataset, but this can be reduced to a minimum.
Make sure you have all the required information and then simply create a new data source name:
- Name of the server
- Name of the database
- Service-Account and of course the password
- Users of the data source
I would recommend this approach, and prepare the workspace owners to change the connection accordingly using the new data source name, this will reduce the downtime (the unavailability) for directquery / live connection or dataset refreshes to a minimum, if this switch will happen between two data refreshes or during an off-time.
Hopefully, this provides some ideas on how to tackle your challenge.
Regards,
Tom
- gaupop135 years agoFrequent Visitor
Thanks TomMartens . I will try to implement and see how it goes.