Lack of correlator in retries of rollbacked transactions (Cygnus HDFS)

[smartcitydevops]

When there is a transaction in Cygnus HDFS with a final message of a rollback

11/28/194:08:45.810 PM

time=2019-11-28T15:08:45.810Z | lvl=INFO | corr=f717a80a-11f0-11ea-a16b-fa163e83ea20 | trans=9a36df35-2d12-49b3-aad5-7f9d2e4c7772 | srv=sc_vlci | subsrv=/sonometros | op=doRollback | comp=Cygnus | msg=com.telefonica.iot.cygnus.sinks.NGSISink[632] : Rollbacking (f49198ca-11f0-11ea-84e4-fa163e83ea20,f717a80a-11f0-11ea-a16b-fa163e83ea20), 10 retries will be done host = pro2-iot-hadoop-nn-01-vlci source = /var/log/cygnus/cygnus_hdfs.log

If we try to check the following retries, the correlator string only is found in a final message of a retry, but , this new transaction lacks its own correlator string

11/28/194:08:56.460 PM | time=2019-11-28T15:08:56.460Z | lvl=INFO | corr=N/A | trans=N/A | srv=N/A | subsrv=N/A | op=processRollbackedBatches | comp=Cygnus | msg=com.telefonica.iot.cygnus.sinks.NGSISink[416] : Finishing internal transaction (f49198ca-11f0-11ea-84e4-fa163e83ea20,f717a80a-11f0-11ea-a16b-fa163e83ea20) host = pro2-iot-hadoop-nn-01-vlci source = /var/log/cygnus/cygnus_hdfs.log
-- | --

We would like to track what happens in this new try, but , with corr=N/A we only can try to guess the logs of this cygnus in a short period of time around this final message. So we understand that this transaction should include its own correlator id

[IvanHdzC]

@smartcitydevops could you provide me a complete log for this scenario where I can fully track the issue in order to propose a solution?

[fgalan]

Implemented by PR #1800