Do Reload target needed after changing any tasks settings or endpoint settings.

[Naga_vivek]

Hi Team,

we are using Qlik replicate tasks we want to verify the scenario where 

if we doing any changes in the Task settings or Endpoint setting(source or target endpoints)

Do those changes reflect in the task by "Resume processing' or again we need to run "Reload Target" to be reflected.

Please let us know what is the best approach to do? what happens in both scenarios(Resume processing and Reload Target ) after any change in task?

Thanks 

N Vivek

[john_wang]

Hello N Vivek, @Naga_vivek ,

Thanks for reaching out to Qlik Community!

In general RESUME a task means starting the task again and read changes from transaction log from the last broken point, the point (stream position) is stored by Qlik Replicate automatically, when the RESUME button is pressed, Replicate start from the stream position. In some scenarios, if the source side or target side cannot meet the synchronization qualification (eg a slot was deleted from PostgreSQL source) then it's meaningless to continue the CDC, a reload is needed. Or startup the task from a given timestamp/SCN/LSN/Pos etc.

The Full Load , or RELOAD (if the task set to Full Load + CDC), Qlik Replicate will initiate the history data load first, and then continue the CDC Change processing automatically.

Hope this helps.

John.

[Heinvandenheuvel]

Well, it would depend on the specific changes made but in general a resume is all you need, for example after adding a source table, changing a tuning parameter, changing the source poll interval.

If you change a column transformation for one or more specific tables then a resume is still fine as those specific tables will be reloaded.

Now if you change a global transformation like maybe uppercase all target column names, or add a 'last changed' date and time stamp then all tables typically need recreating/reloading and for clarity I would sooner restart with reload in stead of resume.

And there are exceptions. For example in that adding a change_DT or change_username column you may decide to manually add the column on target, filll it with some initial values and use  'run advanced' - start by timestamp tables already loaded.

Hein.

[Naga_vivek]

Thanks @john_wang and @Heinvandenheuvel for the reply.

1)In advanced tab Update Change processing mode (read changes from) :  Online Logs    Only
2)In advanced tab Update Apply TLOG truncation prevention policy every (seconds): 600
3)In advanced tab add Internal Parameter: ignoreTxnCtxValidityCheck (set to True)
4)In advanced tab remove Internal Parameter: Provider : SQL Server Native Client 11.0 if present.
5)In advanced tab remove Internal Parameter: safeguardPolicyDesignator : none
6)For the above tasks double click on the Target endpoint connection –
    i.In advanced tab remove Internal Parameter: Provider : SQL Server Native Client 11.0 if present.

7)In advanced tab add Internal Parameter:
i.cdcTimeout (default 600): 6000
ii.executeTimeout (default 60) :600
iii.loadTimeout (default 1200).: 12000

These were some of the changes we are going to do to our replicate task so will the task need Reload target or willl Resume Processing should be fine.

If we are not doing Reload Target while updating the above changes will it affect the task in any way.

If so, what is the parameters that can affect if the reload Target is not performed.

Thanks 

N Vivek.

[Heinvandenheuvel]

Those are all 'details' - no reload needed.

[Naga_vivek]

Hi @john_wang ,

Could you please let me know how the Qlik behaves on this scenario.

If we are stopping any Qlik replicate task and starting after sometime. Whatever changes that are done on Database side when tasks is stopped will that be captured and updated once we resume the task(resume Processing).

Could you please let me know will those changes are captured or not and gets updated in the target?

Thanks 

N Vivek

[john_wang]

Hello @Naga_vivek ,

In general resume Processing works, however that depends on if the source database transaction logs (redo in Oracle, TLOG in SQL Server, Journal in DB2i etc) are available or not. If the transaction logs are truncated (and no way to restore), then resume cannot find the broken point and report error. If the transaction logs are deleted, then need to restore them to the original position.

Regards,

John.

[SushilKumar]

Hello @Naga_vivek 

Reload and resume are Suggested on the basis of issue detected or concluded post log analysis.

in most of scenarios resume work post apply the POA into consideration.

Regards,

Sushil Kumar