We need to improve how we can interact with the TransactionManager which would otherwise timeout the thread fetching the primary ids.
Some discussion started on:
Are there any news on this?
In our business application we increased the abandoned connection timeout from 5' to 1hr to avoid mass indexer connection timeouts.
The Problem is that the indexing progress could durate more than an hour if there is a lot of data in the database.
And also, a one hour timeout value is absolutely insane. It should not be higher that 5'.
Could you tell us please, when a fix for this problem will be available (or you expect it to be)?