camel-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Siano, Stephan" <>
Subject RE: Data Corruption in SFTP in Parallel Multicast branches
Date Fri, 16 Jan 2015 09:01:50 GMT
Hi Lakshmi,

Sure, the exchange should be copied (and it actually is copied). The question is whether also
to do a deep copy of the StreamCache instance (including the underlying file) or modifying
the StreamCache implementations that they do not implement InputStream anymore but return
a new InputStream instance for each consumer (which then work on the same underlying data).
Maybe there is also some third way possible (like having some partial clone of the Stream
cache that does not copy the underlying data, but in that case it might be rather tricky to
determine when an underlying file might be deleted.

All this is somewhat intrusive into the core architecture of Camel, so I wonder what the Camel
architects think about this.

Best regards

-----Original Message-----
From: lakshmi.prashant [] 
Sent: Freitag, 16. Januar 2015 08:47
Subject: RE: Data Corruption in SFTP in Parallel Multicast branches

Hi Stephan,

   The body of the main exchange should be copied to the branch exchanges,
as intended (Option 2 suggested by you).
   But I am not sure if it will lead to performance / memory issues, if
there are more branches with huge data in the body of the main route.


View this message in context:
Sent from the Camel - Users mailing list archive at

View raw message