Problem

From time to time we have to reboot t5memory.

Sometime that may happen in active time and cause some queries and processes to fail.

To avoid that we already have re-try logic for import, export and reorganise.

Currently when re-import fails and is retried, the failing is logged as error, despite later it succeeds.

Solution

Implement same principles for update operations.

Re-import failing if it succeeds automatically later should only be logged as info.