Hello. Questions for the experts.
Transaction logs have three performance options:
1 - Favor runtime
2 - Standard
3 - Favor restart recovery time
Each type of performance has its own values for creating checkpoints and redo. This can be verified using the “Database.RM.sys” statistics:
1 - Transactional logging = Favor runtime:
Database.RM.Sys.Log.Type = Circular
Database.RM.Sys.M.Chkpt.Interval = 85
Database.RM.Sys.M.Log.Size = 512
Database.RM.Sys.M.Redo.Limit = 255
Database.RM.Sys.Phase = Normal
2 - Transactional logging = Standard:
Database.RM.Sys.Log.Type = Circular
Database.RM.Sys.M.Chkpt.Interval = 40
Database.RM.Sys.M.Log.Size = 512
Database.RM.Sys.M.Redo.Limit = 120
Database.RM.Sys.Phase = Normal
3 - Transactional logging = Favor restart recovery time:
Database.RM.Sys.Log.Type = Circular
Database.RM.Sys.M.Chkpt.Interval = 8
Database.RM.Sys.M.Log.Size = 512
Database.RM.Sys.M.Redo.Limit = 24
Database.RM.Sys.Phase = Normal
There is an article that describes the difference: Transactional Logging and how it operates - Customer Support
But after reading the article, some questions remain:
1 - What is the difference between Chkpt.Interval and Redo.Limit?
2 - Are the sizes of Chkpt.Interval and Redo.Limit values for each database or values for all databases on the server?
3 - How do the values of Chkpt.Interval and Redo.Limit interact?
4 - How do Log.Size, Chkpt.Interval, and Redo.Limit relate to each other?
5 - Am I correct in understanding that Chkpt.Interval is the size of transaction logs required for recording in UBM?
6 - If point 5 is correct, what happens when Redo.Limit reaches its limit?
7 - What happens when Redo.Limit reaches its limit, but the server cannot perform Flush?
P.S. If you write a response, please provide as much detail as possible, as there are no answers to these questions in the articles. Thank you.