Self-repair of a data source marked as unavailable

Manish shared this idea 8 months ago
Completed

There are cases where a data source get's marked as unavailable which then requires manual intervention, in the form of 'test connection' via the UI, or running a web service to refresh the source.


How can we have this as part of a self-repair option, so that once a data source has been marked as unavailable, the data source can attempt to repair itself. This is extremely important in cloud offerings where down-time needs to be kept at a minimum.

Comments (2)

photo
1

Hey Manish,


I've logged this request on your behalf with some ideas, what I'm thinking is a way to configure the data source 'refresh' web-service call on a schedule. So you could say, if data source fails, then run refresh call, then if fails, wait x mins, run again, etc..

Then once it get's to it's final attempt and still fails, to email YF admin. However in terms of the best way to implement I'll leave that for product team review and circle back with more details (once reviewed).

If you have any ideas happy to add also.


Thanks,

David

photo
1

Hi Manish,

just letting you know this item has now been implemented in the latest 8.0.10 release . In order to take advantage of this change, please ensure you insert the following into the Yellowfin repository DB;

insert into configuration
values (1,'SYSTEM','NEVERMARKUNAVAILABLE', 'TRUE')
By inserting the above, the data source will never be flagged as unavailable. This means even if the data source connection cannot be made during connection pool creation, it will still be available for future connection attempts, without needing to manually test the data source or web services.


Please let mw know if you have any questions on this.

Regards,

David