ZIP aggregation and persistence

[Stuart Drapier]

Hi,

I'm using Camel to read files from directories on a filesystem, aggregate to a ZIP file and forward to AMQ. This all works brilliantly until the process is unexpectedly shutdown - we lose any pending files in the aggregator.

I understand this is because the aggregator by default uses the MemoryAggregationRepository and this is or course volatile. I need to ensure that the files aren't lost in this failure scenario. I've tried both the LevelDB and HawtDB repository implementations and both don't support java.io.File body types; are there any options I am missing? I'm happy for the files to exist on disk whilst they are aggregated.

Any help much appreciated,

I'm aware this question was asked a few years ago but I didn't want to resurrect a stale thread:
http://camel.465427.n5.nabble.com/ZipAggregationStrategy-with-persistent-store-tt5782458.html

[Claus Ibsen]

Hi

What is your completion criterias for the aggregator?

If possible you can maybe also try to only consume the files from the
file endpoint if you have all needed files, so they stay there on disk
until you have all of them.


On Mon, Aug 20, 2018 at 3:26 PM, Stuart Drapier

Post by Stuart Drapier
Hi,
I'm using Camel to read files from directories on a filesystem, aggregate to a ZIP file and forward to AMQ. This all works brilliantly until the process is unexpectedly shutdown - we lose any pending files in the aggregator.
I understand this is because the aggregator by default uses the MemoryAggregationRepository and this is or course volatile. I need to ensure that the files aren't lost in this failure scenario. I've tried both the LevelDB and HawtDB repository implementations and both don't support java.io.File body types; are there any options I am missing? I'm happy for the files to exist on disk whilst they are aggregated.
Any help much appreciated,
http://camel.465427.n5.nabble.com/ZipAggregationStrategy-with-persistent-store-tt5782458.html