attreps_status vs attrep_txn_status table

[MoeyE]

Hi Team,

Apologies for my countless questions. I recently came across a case where we need to find the best way to continue a task after a failover event and it got me thinking.

I know that the attrep_status table can take minutes to be updated and therefore it could cause duplicates. So it's probably not the best solution (target doesn't detect duplicates) . So I thought that attrep_txn_status table could be used to get a more current checkpoint of the status?

1. Can the CHECKPOINT value be used in the case of failover?

2. How often does the CHECKPOINT column in the attrep_txn_status table get updated?

3. Does it represent the most current point of the task transactions?

4. What exactly is the checkpoint? Is it like the source_tail_position from the attrep_status table? streamPosition? SCN? etc...

Any guidance on this would be great. Thanks.

Kind Regards,

Mohammed 

[Heinvandenheuvel]

>> I know that the attrep_status table can take minutes to be updated and therefore it could cause duplicates. 

E.2 Replication status
The attrep_status table contains the current status of each replication task and the target data.  Although updates for the tasks in the attrep_status table are generated every few seconds, Replicate will only apply the updates after it has applied changes to the target tables. In some cases, this may take a few minutes.

It may take minutes if the target is slow taking TARGET_APPLY thread commands, in which case everything going to the target is slow.

>> 2. How often does the CHECKPOINT column in the attrep_txn_status table get updated?

The name of the table is actually attrep_txn_state. It is an undocumented internal table processed with undocumented internal algoritme which may change with every release. The old sales rule applies: If you need to ask the price (workings) then you cannot afford (to use) it.

Stick to the well (well.... after several re-reads)  documented attrep_status.  

Good luck,

Hein.

[MoeyE]

Hi Hein,

Thanks for the answer, and the wisdom!

Kind Regards,

Mohammed

[MoeyE]

Hi Hein,

Again I have one more question for now. How exactly does Replicate keep track of the task status? I assume that it stores it's own attrep_status details or some similar details somewhere on the Replicate server. However I don't know the specifcs. Thanks.

Edit:

Never mind I think I found the answer here https://help.qlik.com/en-US/replicate/May2023/Content/Global_Common/Content/SharedEMReplicate/Tasks%...

Kind regards,

Mohammed