2004-12-13 jrandom

* Added some error checking on the new client send job (thanks duck!)
    * Implemented tunnel rejection based on bandwidth usage (rejecting tunnels
      proportional to the bytes allocated in existing tunnels vs the bytes
      allowed through the bandwidth limiter).
    * Enable a new configuration parameter for triggering a tunnel rebuild
      (tunnel.maxTunnelFailures), where that is the max allowed test failures
      before killing the tunnel (default 0).
    * Gather more data that we rank capacity by (now we monitor and balance the
      data from 10m/30m/60m/1d instead of just 10m/60m/1d).
    * Fix a truncation/type conversion problem on the long term capacity
      values (we were ignoring the daily stats outright)
This commit is contained in:
jrandom
2004-12-13 13:45:52 +00:00
committed by zzz
parent 83c6eac017
commit 9e16bc203a
7 changed files with 167 additions and 25 deletions

View File

@ -1,4 +1,17 @@
$Id: history.txt,v 1.103 2004/12/11 02:05:12 jrandom Exp $
$Id: history.txt,v 1.104 2004/12/11 04:26:24 jrandom Exp $
2004-12-13 jrandom
* Added some error checking on the new client send job (thanks duck!)
* Implemented tunnel rejection based on bandwidth usage (rejecting tunnels
proportional to the bytes allocated in existing tunnels vs the bytes
allowed through the bandwidth limiter).
* Enable a new configuration parameter for triggering a tunnel rebuild
(tunnel.maxTunnelFailures), where that is the max allowed test failures
before killing the tunnel (default 0).
* Gather more data that we rank capacity by (now we monitor and balance the
data from 10m/30m/60m/1d instead of just 10m/60m/1d).
* Fix a truncation/type conversion problem on the long term capacity
values (we were ignoring the daily stats outright)
2004-12-11 jrandom
* Fix the missing HTTP timeout, which was caused by the deferred syn used