Feature #2580
Updated by Tom Clegg about 10 years ago
Ensure Perform some simple benchmarking for Keep, and (if time permits) compare new Keep behaves well under heavy load. benchmarks with old Keep. * Start keep server. (Use "\time keepd" so we get a "maxresident" memory footprint data point.) * Run Use side by side results from three concurrent "ab" processes (apt-get install apache2-utils) -- one read, one write, one bogus -- with lots of requests and lots of concurrency. * Same test with SSL. * Inspect data volumes after test in case it corrupted data, left tmp files lying around unexpectedly, etc. (No need to automate this for now.) * Paste "ab" output (and "time" statistics) on a wiki page. bogus.