Hi All,
We have more than 100 jobs right now in Databricks which connect to on-prem database like Oracle. Connection to oracle is made through notebooks using jdbc thin client and using com.oracle.ojdbc:ojdbc10:19.3.0.0 or com.oracle.ojdbc:ojdbc8:19.3.0.0. The jobs were all running successfully since 2 months until yesterday, but suddenly we started getting error from today: The Network Adaptor could not establish the connection. On seeing the log, we get error: No route to host, socket connect lapse 17s. The issue is intermittent where some tasks are getting successful while others are getting failed. All tasks are connecting to same Oracle server and same database/schema. We have a hub and spoke vnet and hub is connected to on-prem databases via site-to-site VPN.
What could be the possible reasons of this failure?
