On Tue, Jan 5, 2016 at 12:09 AM, Rob Miller <[email protected]> wrote: > I could maybe be convinced, but I'm leaning negative on this one. > Logstreamer is already a pretty complicated feature, with a lot of fiddly > code. I'm loathe to add yet more features / complexity unless the win is > big. It doesn't seem to terribly unreasonable for users to create an empty > folder on the machine in question.
Thank you Rob for your response. Have you looked at my patch [1]? From my perspective the patch brings no additional complexity to Logstreamer. It just moves two function calls from the NewLogstreamSet function to the ScanForLogstreams function. And I think it does bring flexibility to users. In my case, to reduce configuration management complexity, I'd like to deploy the same Heka configuration on all the nodes of my data center. For this to work I'd need Logstreamer to be adaptive, and to not fail when the log_directory doesn't exist. After all Logstreamer already doesn't fail when there are no files matching file_match, so I don't quite understand why it should fail when log_directory does not exist. Please tell me if I'm not viewing this from the proper angle. Thanks! [1] <https://github.com/mozilla-services/heka/compare/dev...elemoine:logstreamer> _______________________________________________ Heka mailing list [email protected] https://mail.mozilla.org/listinfo/heka
