Missing Rows after tSalesforceOutputBulkExec

[Anonymous]

Hi - I'm in the process of removing a paid package from a Salesforce env and as part of that process, I've created a new object and need to transfer data from the existing, paid-for, object to the new object that we've created ourselves.

There are around 1.3m records that need to be created.

I'm able to process all 1.3m records right up to tSalesforceOutputBulkExec - tSalesforceOutputBulkExec creates the bulk .csv file with all 1.3m rows but the logs show that only 330,275‬ rows are processed.

I'm pretty new to Talend (this is my first talend job) and I'm struggling to understand why the remaining records arent being processed.

Any help would be very much appreciated

Thanks

[TRF]

Can you see any error on salesforce side?
For example, if you're on a sandbox, are you able to inject so much records, or maybe your datastorage is near of 100%.
You may also have lost the connection due to the elapse time to process these records.
Many tracks to explore.
What if you limit to 10 or 25% of the expected records?

[Anonymous]

Hey - Thanks so much for the quick reply.

I'm running the job in a full-copy Sandbox so should have no limit on the number of records I can create as long as it's within data storage/api limits. There's about 10GB worth of available space in the SF env and the number of records created should come to about 2.6GB of data storage.

I've just re-run the job with about 10% (144,540) of the required records and the same thing has happened. Only about 30% of the rows make it through to the logs. (screenshot attached) and the other 70% go "missing". (that's a similar ratio to when I ran with 1.3m records).

I've also provided a shot of the advanced settings on the tSalesforceOutputBulkExec component.

Also is there a way to track the progress of the tSalesforceOutputBulkExec job other than waiting for the logs to be created at the end? Like a real-time view of how many records are being created in the env per second?

Thanks again for the help.

M

Capture2.PNG
Capture3.PNG

[TRF]

Strange behaviour. I never saw that before and I've practice a lot.
To track how your job works on salesforce side, go to Setup, enter Bulk Data in the Quick Find box, then select Bulk Data Load Jobs. Click the Job ID to view the job detail page, as explained here https://developer.salesforce.com/docs/atlas.en-us.api_asynch.meta/api_asynch/asynch_api_batches_moni...

[Anonymous]

Thanks for directing me to the bulk data load section in SF.

From there I was able to see that the missing 70% records were in batches that failed because they were unable to obtain exclusive access to a particular record.

TooManyLockFailure : Too many lock failure 200 Trying again later.;

It tried a bunch of times and then times out because it was never able to get exclusive access.

Now I just need to figure out how I can avoid the above locking error. I've too many records to be able to run the concurrency mode as 'Serial' - I'll start a new thread for that if I'm not able to locate an answer elsewhere in the forum.

Thanks again for your help

[TRF]

Vive more details.
Are you upserting records?
Do you have lookups to establish during the operation?
Are you using external ids for that or ids?
Do you have events triggering on salesforce side during the operation?

[TRF]

Did this help? If so, thanks to mark your case as solved (Kudo also accepted)