And it ignores the rest. No adapter/broker/unix/any errors. each line is has the same length and i put 128M for the adapter heap memory. The file only has 2k lines.
has anyone encountered this before?
I have already reported this to wM but they have not encountered this problem before.
Rob, it is a typical IO adapter process, using file poll to read the file and using the delimited output to parse the content of the file and publish document per line. I have tripled check the record layout and it is very consistent. At the end I reprocessed the same file and did not have any problem at all, every line was processed correctly. Weird!
We are having almost same problems as what LS indicated. We have raised the issue with wM. This happens with IO Adapter 4.2 version and happens only with FilePoll Operation.
We are able to simulate the porblem in 2-3 workstations at our site.
Our problem is there are more than 6 files coming at the same time and each having more than 4000+ records. FilePoll picks 2-3 of say 6 files, processes half of each, and moves all files to completion directory.