On Dec 10, 2007, at 1:10 PM, Matthew Toseland wrote:

IMHO the next step forward is simply to log location changes and display them 
either on the location page or on a subpage, or as CSV data (or perhaps 
through SNMP) so it can be graphed externally. Maybe for the node's peers as 
well as itself. Are you interested in doing some data collection code? Lets 
discover whether there actually is a problem with location drift before we 
try to solve it ...

I think that it should be observable with a few more datastore stats. Specifically the datastore location, and the location of succeeding requests.

A quick implementation is attached. It only accounts for what is in the datastore since-launch (but shows a percent thereto), and takes up several megabytes more of memory to track the location of the datastore/cache in a running average.

If we are leaving our stores behind, I would predict that for many nodes (running suffecient to have meaningful stats) that the 'storeDist' value may become arbitrary. And if it approaches the 'furthestSuccess' value the store is "left-behind"; whereas the cacheDist will maintain approx. the same (once reliable).

Do you want this in the trunk? I don't think that I have the time to do much data collection.

--
Robert Hailey

Attachment: freenet-datastore-location-stats.diff
Description: Binary data



_______________________________________________
Devl mailing list
[email protected]
http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to