Hrm. In our terminology, what you'd be doing is "writing an adaptor", or "modifying the agent" -- you shouldn't need to change any significant part of the existing agent code.
There are already adaptors for pulling data out of files, without parsing. They come in several flavors, depending on how you want the adaptor to behave if the file is modified and whether you want line buffering. Have a look at the FileTailingAdaptor and subclasses for examples and a place to start. See https://svn.apache.org/repos/asf/incubator/chukwa/trunk/src/main/java/org/apache/hadoop/chukwa/datacollection/adaptor/filetailer/FileTailingAdaptor.java --Ari On Mon, Feb 4, 2013 at 4:24 AM, Farrokh Shahriari <[email protected]> wrote: > Hi > I've read in the Chukwa's site : > http://incubator.apache.org/chukwa/docs/r0.5.0/agent.html , that we can > develop our agent's code by defining new adaptors. As there is no example > for writing custom agent, I'd be glad if someone can help me. > My agent is so simpe,it should read a file or folder and send a line of that > to collector for inserting to HBase without doing any operation like > parsing. > > Tnx -- Ari Rabkin [email protected] Princeton Computer Science Department
