As per my knowledge when you set the online logs only --> Replicate will read the data from the TLOGS.
Backup logs --> Replicate will read the data from the backups (in the disk) that is maintained by the DBA. And it will read in the regular interval (Set by DBA at their end), which may lead to latency.
Please correct me if am wrong.
And also could you please let me know the difference between online log and archival backup log in oracle and how it works with respect to replicate.
"when you set the online logs only --> Replicate will read the data from the TLOGS."
Correct. When this option is selected, Qlik Replicate will try and find the changes in the online transaction logs only.
"Backup logs --> Replicate will read the data from the backups (in the disk) that is maintained by the DBA. And it will read in the regular interval (Set by DBA at their end), which may lead to latency. "
Yes also correct. Qlik Replicate will try and find the changes in the backup transaction logs only. Selecting this method results in increased latency due to the interval between backups. The actual latency time will remain constant, but will vary according to the backup schedule.
For Oracle:
"the difference between online log and archival backup log in oracle and how it works with respect to replicate."
Hi, thank you for the post to the QDI Forums. For SQL Server Endpoint depending on the settings on the Endpoint Replicate will read and can read from the online TLOG and the backups as long as the Backups are defined to the internal MSDB DB of the SQL Server you are connecting too. If you decide to use Backups Only you would want to specify the backup file location and the file level access to the path defined for optimal performance.
The Oracle setup of the Online Redo Logs and the Archived logs is also dependent how you setup the Source Endpoint for Oracle would be same type of setup and depending on the Oracle DB configuration this would be how Replicate would access the Archive Redo Logs via the location setup in this Source environment. You can also check the below link on the settings per your requirements. Hope this information is helpful and please let us know.
@suvbin I've seen many questions from you for many topics, which is great. It suggests that you work with the product a lot and that it would be worth your while to try to figure out some nitty-gritties on exactly how it works. The REPTASK log files potentially have a wealth of information notably in 'debug' mode (verbose is often too much).
To better understand how this all works you may want to spend some time to study those logs for SIMPLE tasks with low rowcount and low change rate.
For the actual question on hand you will understand better how Replicate works by studying the following simplified version of the queries it uses:
SQL server:
select bs.backup_start_date, bs.first_lsn,
bs.backup_size/(1024*1024) "MB", bmf.physical_device_name
from msdb.dbo.backupmediafamily bmf, msdb.dbo.backupset bs
where bmf.media_set_id = bs.media_set_id
and bs.database_name=db_name() and bs.type='L'
and bs.backup_start_date > getdate() - 2
Oracle:
select SEQUENCE#, thread#, first_change#, blocks, name
from v$archived_log
where first_time IS NOT NULL and name IS NOT NULL
and status = 'A' and next_time >= sysdate - 2;
And yes, archive log only, will create 'peaky' delays as the data is only available once the backup is made which is for example every 15 minutes, or every time a REDOLOG (1GB? 10GB?) is filled or actively switched for Oracle.
.png "Former Employee"){kind=icon}
