ORACLE CDC ISSUE

[msantiago]

Hello Everyone,

We are new to Qlik, and we are using Qlik enterprise manager. Currently we are experiencing CDC issue with Oracle as a source and as comparison we also have a similar connection using Oracle as well. 

Appreciate the response 😁

[SachinB]

Hello @msantiago ,

Thank you for reaching out to the Qlik community!

Both the tasks are connected to same source endpoint and same db? and both are parent log stream tasks ?

One is appearing to be zero latency and other parent task is having the latency around 58 minutes.

Could you please check, do you have any mass updates from source? Also check is it source or Target latency in change processing tab.

Please check target endpoint is able to access or not.

Please set source capture, Target apply, Sorter and Performance in Trace 15 to 20 minutes.
Above parameters will help us to identify the bottle neck of the issue, we would request you to create case to check further on latency.

However we do have below article which will help you to understand more in details.

https://community.qlik.com/t5/Official-Support-Articles/Troubleshooting-Qlik-Replicate-Latency/ta-p/...

Regards,

Sachin B

[msantiago]

Hi @SachinB,

Thank you for your response, already checked the youtube video for this topic and this helped me on the other Qlik tasks and it worked.
However for this particular case, this is really weird in a way that all parameters mentioned in the video was applied. Will now
follow your advise for the logs.

Thank you!

[Heinvandenheuvel]

Collect more details from the teptask log, with TARGET_APPLY set to DEBUG or even VERBOSE.

If you cannot figure it out from there you may want to open a fresh discussion in https://community.qlik.com/t5/Qlik-Replicate/bd-p/qlik-replicate-discussions . Please consider attaching that reptask log, or at least a significant chunk from it, to the new topic as text to help folks help you.

The Enterprise Manager is just 'the messenger' reporting on Replicate, but it does not have control itself.

Hein.

[msantiago]

Hi @SachinB  / @Heinvandenheuvel 

Here is the log that is set to Trace. Let me know if this is sufficient enough data. Hope you can help me out.

Thanks,

Myk. 

[Heinvandenheuvel]

Nasty  Log.  Wrong forum.

SORTER information is too hard to deal with. Who asked for that sh*t?

The connection to the source is terrible. Have network and/or oracle folks lok at that with 'ping', TNSping an more.

>>> 2023-09-28T22:47:32 [SOURCE_CAPTURE ]I: Trying to reconnect to the source database

>>> 2023-09-28T23:53:22 [SOURCE_CAPTURE ]I: The connection to the source database reestablished

More than 5 minutes doing nothing.

>>> 2023-09-29T08:41:09 [PERFORMANCE ]T: Completed to read from archived Redo log 51200000 bytes at offset 00000000155cb600 with rc 1, read time is 16927 ms

And a handful like it. 16 seconds for 50 MB = 3MB/second. Using a bicycle? How much data is produced per second (calculate using Archive Log sizes + times).  Seems a good reason to fall behind.

Get a better log - NO SORTER, run it longer, and move over to the Replicate Forum instead of AEM.

See you there.

[msantiago]

Hi @Heinvandenheuvel ,

Ok will do as advised. will get to that forum