Below are some of the Frequently Asked Questions by customers , while installing and maintaining the Dbvisit replicate.
1. Why is PK supplemental logging enabled at DB(Database) level? According the documentation, this is only needed when we replicate DDL.
By default this is enabled to make replication more convenient - to allow majority of users to make changes, add new tables, schemas etc to replicate even after it is up and running. As described in our documentation (https://dbvisit.atlassian.net/wiki/display/ugd7/Enable+supplemental+logging ) The supplemental logging can be enabled at two levels - DATABASE LEVEL and TABLE LEVEL.We have not had any issues in past about the overhead in db when the logging is enabled.
So if you don't want to enable the supplemental logging at DB level, you have to do it at table level and this would need to be done for every table which is part of replication now and any future additions.
2. Why we need to be able to LOCK ANY TABLE in the source DB to replicate DDL? This is a highly elevated privilege to give in PRODUCTION .
The locks are used to get the SCN (during preparing new tables) to avoid any dirty writes - possibly causing an issue during instantiation. You could possibly avoid this by locking table yourself and preparing the table (new tables) using that SCN (obtained while table is locked).
You can prepare table with no ddl if you choose to do so but if you are running any ddl operations (including truncate) you will have to complete the ddl operations on target db yourself.
Get the SCN
LOCK TABLE SCOTT.TEST1 IN SHARE MODE; SELECT CURRENT_SCN FROM V$DATABASE; ROLLBACK; /
Prepare table as of SCN noddl
dbvrep> PREPARE as of 79728416 TABLE scott.test1 noddl
Bottom line is you should be able to secure your database and Dbvisit replicate can handle that but it will require complete understanding of what the consequence of doing so is. We recommend testing these in your sandbox environment to ensure there are no unexpected outcomes.
3. Can Dbvisit Replicate support partitioned table with TDE ?
We do support partitioned tables.
We don't currently support TDE (though this is on the road map for the very near future!).
4. What type of encryption is used on PLOG files?
This is more network related than file based - and you can find more information on this here:
5. What is the impact on redolog since MINE process is reading from redolog?
It does not impact directly on redolog processing - as all we are doing is reading from the redolog as soon as it has been written to (think of like tailing a file).
The gathering of the redo information (Fetcher) and/or processing it (MINE) are processes which may take up some CPU on the source system - but that is just overall resource consumption on that server - and it does not directly impact upon the redo log writing.
6. Any administration GUI console for Dbvisit Replicate?
We don't have a GUI console yet in replicate,but we have a command console:
7. How to schedule the alerts for DB sync.
8. How fast is the sync between source and target? (Eg: If we generate 100MB of archivelog per Minute).
This is very difficult to answer generically as "it all depends" on a number of factors:
ii) Type of data being replicated
ii) Volume of data being replicated (ie: no. of tables/schemas in a replication and the changes being made in their on source), not everything in the redo log from source will be pushed to the target with a logical replication solution.
We do ask all the customers to run a pre-flight data discovery scripts which would give a general idea about the load of the system and data.