Hi Aleksey Zhivotchenko
Alright this case looks specific. It may be a bug with OID or Oracle DB. Raise with Oracle Because it's hanging issue which can happen due to config issues or underlying bug.
check under Logs if you see Pointers
We are currently in upgrade mode from 11g stack to 12c stack.
On the 11g stack: oid is mainly managed by the OPMN processors.
On the 12c stack: oid is a component type.
We found a major difference between the 2 versions when it comes to handling outages of the Database that OID is connected to. In 11g -> the opmn processors just die and you will need to start them again if the Database goes down for maintenance. In 12c -> the component just hangs, in some cases occupying 1 full CPU thread completely and showing high CPU.
Question: What experience or suggested can you give in this case?
We fully understand that Database can die or go down or something of that sort: but we want to monitor this whole thing. We created the monitoring for viewing the status of the components in 12c but found out that it does not work since the component does not die but just hangs. In addition to that: we don't even understand why it needs to die, why can't it simply reconnect to the Oracle Database in the same matter as Data Source reconnect in the weblogic.
Question 2: do you know of any configuration changes in 12c that can at least bring similar functionality in 12c to 11g? So that at least 12c component will simply shut itself down if the database servicing it will go down as well.