Error using File Channel - LFA-E-SRCMSS, Missing source file

[Abrie_M]

Good day, we are doing a full load using file channel target endpoint. 

The following error is displayed in the logs multiple times - seems the full load is stuck there:

00004716: 2023-08-03T13:53:34:816232 [FILE_TRANSFER ]T: last cifta error status = 778778778, error message = <LFA-E-SRCMSS, Missing source file <ID:37|PATH:C:\Program Files\Attunity\Replicate\data\databases\Sanlam\fc\streams\cdc\0000000D.fcd>> (at_cifta.c:2879)
00004716: 2023-08-03T13:53:34:816232 [FILE_TRANSFER ]T: last application status = 3 (at_cifta.c:2880)
00004716: 2023-08-03T13:53:34:816232 [FILE_TRANSFER ]T: last callback error status = 0, last callback called = <file_error> (at_cifta.c:2881)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]V: (COMMON->ciftac_query_state_fields( cifta_client_object->ciftac, CIFTAC_STF_ALL, &state )) (at_cifta.c:2562)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]V: (COMMON->ciftac_query_state_fields( cifta_client_object->ciftac, CIFTAC_STF_ALL, &state )) (at_cifta.c:2718)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: cifta status is: up = 1, number of pending files = 65 (at_cifta.c:2719)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: last callback method name is <file_error>, last error message is <LFA-E-SRCMSS, Missing source file <ID:37|PATH:C:\Program Files\Attunity\Replicate\data\databases\Sanlam\fc\streams\cdc\0000000D.fcd>> (at_cifta.c:2730)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: last successfully transferred file: id = 36, source = <C:\Program Files\Attunity\Replicate\data\databases\Sanlam\fc\streams\cdc\0000000C.fcd> (at_cifta.c:2739)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: consequent_connection_errors_counter = 0 (at_cifta.c:2785)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: internal_error_timestamp_array counter = 0 (at_cifta.c:2786)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: transfer_error_timestamp_array counter = 3 (at_cifta.c:2787)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: last times: transfer = 1691069825836382, internal error = 0, transfer error = 1691069841529476 (at_cifta.c:2806)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: cifta status is: running_status=<FILE_ERROR_RETRYING>, number_of_pending_files = 65 (at_cifta.c:2867)
00004716: 2023-08-03T13:53:35:826183 [FILE_TRANSFER ]T: cifta status: last successfully transferred id is 36 <C:\Program Files\Attunity\Replicate\data\databases\Sanlam\fc\streams\cdc\0000000C.fcd> (at_cifta.c:2872)

Any suggestions as where to start looking for the cause?

Thanx a lot!

Abrie

[john_wang]

Hello @Abrie_M ,

Thanks for opening the article!

In a typical configuration involving a File Channel, we have 2 Replicate server machines, a local server which runs the task that reads from the source database, saves the data into *.fcd files, and transfers the files across the network to the remote server, where another task would read the *.fcd files and write to the target, the file transfer mechanism is the File Channel. Is this your configuration? If so then there was probably a network issue between the 2 Replicate server machines which caused the file transfer to fail. You may check to see if the file "0000000D.fcd" exists in OS level or not.

When there is such error we cannot resume the function ,  please take note the limitations in user guide:

• You cannot use the Full Load resume function if you are using the File Channel endpoint. To resume a
  Full Load operation, you must delete the original data and then run the task again.

Hope this helps.

Regards,

John.

[Heinvandenheuvel]

Hmmm, 

This can be tricky. You are showing part of the target task log. How does does Source task looks like when you grep (search) for 0000000C.fcd  and 0000000D.fcd. Did either side run out of disk space? Carefully study the directory listings for \Sanlam\fc\streams\cdc\ on both sides.

According to https://www.epochconverter.com/     the last transfer time was

1691069825836382  GMT: Thursday, August 3, 2023 13:37:05.836, and the last error time was 

1691069841529476  GMT: Thursday, August 3, 2023 13:37:21.529 

You are on file 12 out of 65.  

I suspect with @john_wang that the most effective coarse of action is to delete the \Sanlam\fc\streams\cdc\ directories and everything under it and restart the load. If you have the space, instead of delete rename the directories to FC_FAIL_1 or similar to study and/or compare with the next run.

Do you have autodelete enabled?

Some 10 years ago, I had to help maintain a site which heavily relied on File Channel - with no autodelete. To help maage that I wrote two PERL scripts at the time. Maybe the scripts themselves or the concepts in it are of use to you to see what it happening? Admittedly Perl is easier to write than to read :-). Attached as .TXT

Good luck,

Hein

[Abrie_M]

Thanx John and Hein for your replies. I'll play around and see what we can figure out.