Hi Folks,

I'd like to report an issue we have with Attunity Replicate. We are using an instance from the AWS Marketplace (/w supplied hourly licence) per:
Code:
00003576: 2018-01-19T02:50:08 [AT_GLOBAL       ]I:  Task Server Log (V6.0.0.238 EC2AMAZ-ALOK7HJ Microsoft Windows Server 2012  (build 9200) 64-bit, PID: 2940) started at Fri Jan 19 02:50:08 2018  (at_logger.c:2470)
00003576: 2018-01-19T02:50:08 [AT_GLOBAL       ]I:  Licensed to AWS Marketplace - Attunity Replicate Hourly, permanent license, sources: (Oracle,SQLServer,MySQL,PostgreSQL,DB2LUW), targets: (Oracle,SQLServer,MySQL,PostgreSQL,Teradata,Redshift,Hadoop), all hosts  (at_logger.c:2473)
We are attempting to replicate data from a MySQL RDS instance to a Redshift instance. When we setup the job, the full load works fine, it reports finding:
Code:
00006920: 2018-01-18T10:34:31 [SOURCE_CAPTURE  ]I:  Set position to initial context 'now'  (mysql_endpoint_capture.c:3080)
00006920: 2018-01-18T10:34:31 [SOURCE_CAPTURE  ]I:  Setting position in binlog 'mysql-bin-changelog.009678' at 9178452  (mysql_endpoint_capture.c:785)
at the end of the full load, it notes:
Code:
00006920: 2018-01-19T02:45:04 [SOURCE_CAPTURE  ]I:  > ROTATE_EVENT  (mysql_endpoint_capture.c:2959)
00004872: 2018-01-19T02:46:08 [SORTER          ]I:  Final saved task state. Stream position mysql-bin-changelog.009874Am:143093:-1:143124:42408507220456:mysql-bin-changelog.009874Am:142947, Source id 4802602, next Target id 12401841, confirmed Target id 12400325   (sorter.c:655)
00004008: 2018-01-19T02:46:12 [TASK_MANAGER    ]I:  Subtask #0 ended  (replicationtask_util.c:937)
00004008: 2018-01-19T02:46:12 [TASK_MANAGER    ]I:  Task management thread terminated  (replicationtask.c:3105)
00005016: 2018-01-19T02:46:12 [SERVER          ]I:  Client session (ID 7426) closed  (dispatcher.c:200)
00005016: 2018-01-19T02:46:12 [UTILITIES       ]I:  The last state is saved to file 'C:\Program Files\Attunity\Replicate\data\tasks\Afterpay/StateManager/ars_saved_state_000001.sts' at Thu, 18 Jan 2018 15:46:11 GMT (1516290371809835)  (statemanager.c:601)
Notice the additional characters at the end of the binlog filename ? We feel like they could be the root of the problem.

Upon starting the incremental load, we see:
Code:
00000908: 2018-01-19T02:50:09 [SORTER          ]I:  Start the task using saved state. Start source from stream position mysql-bin-changelog.009874Am:143093:-1:143124:42408507220456:mysql-bin-changelog.009874Am:142947 and id 4802602. Confirmed target id is 12400325, next target id is 12401841  (sorter.c:449)
00003900: 2018-01-19T02:50:09 [SOURCE_CAPTURE  ]I:  Resume TABLE_MAP at file 'mysql-bin-changelog.009874Am', pos '142947'  (mysql_endpoint_capture.c:3173)
00003900: 2018-01-19T02:50:09 [SOURCE_CAPTURE  ]I:  Setting position in binlog 'mysql-bin-changelog.009874Am' at 142947  (mysql_endpoint_capture.c:785)
00003900: 2018-01-19T02:50:09 [SOURCE_CAPTURE  ]I:  System var 'binlog_checksum' = 'CRC32'  (mysql_endpoint_capture.c:272)
00003900: 2018-01-19T02:50:09 [SOURCE_CAPTURE  ]I:  Position was set in binlog 'mysql-bin-changelog.009874Am' at 142947  (mysql_endpoint_capture.c:810)
00003900: 2018-01-19T02:50:09 [SOURCE_CAPTURE  ]I:  Error 1236 (Could not find first log file name in binary log index file) reading binlog. Try reconnect  (mysql_endpoint_capture.c:999)
So.... obviously it just fails. We attempted to change to use the official MySQL ODBC drivers, but sadly the AWS Marketplace image isnt licensed for that feature, so we can only use the supplied driver.

The problem is a little hard to visualize in the text above as the additional characters are in fact unicode I believe. If someone from Attunity can supply an email address I'm happy to send in the full logs.

FWIW - We have considered modifying the state file with a hex editor to remove the extra data after the filename, but presumably we would be stuck if we stopped/started the process again as the state file would get overwritten with bad data again.

If anyone has any suggestions of advice, please share.