9218f7b82c
deal with not having tunnels a bit earlier
2004-07-25 23:51:07 +00:00
edaf7aee5d
* for the moment, remove the 'isFailing' check for peers who have failing tunnels
...
(we need a more sophisticated algorithm than the one in place for it to be effective)
* fix for the profileOrganizer to work safely in the sim
2004-07-25 23:46:55 +00:00
43c18d0f4d
(techincally) reduced the minimum bandwidth rate to 1KBps, but NO ONE SHOULD SET IT THAT LOW. do not reduce your limits below 6KBps until More Stuff Gets Done.
...
logging
2004-07-25 23:43:13 +00:00
65d85f7479
the vast majority of messages on the live net are under 2KB
2004-07-25 23:40:08 +00:00
476e23db5b
new stat monitoring the netDb search reply message sizes
2004-07-25 23:35:50 +00:00
ce3e7e623c
handle disconnect while there are still requests pending
2004-07-24 17:54:49 +00:00
f170ae741e
0.3.3, backwards compatible, to be released Real Soon
2004-07-24 02:13:37 +00:00
740a2da702
more consistent html
2004-07-24 01:59:27 +00:00
85c8e56417
fixed a strange bug when the .wait delay is really accurate (too accurrate..). thanks ZeroCool for help debugging this!
2004-07-24 01:10:11 +00:00
9f4439583d
expose some data points for the new console, and cleaned up some html
...
new piece of data exposed and maintained is a list of router contexts - shown as a singleton off RouterContext - allowing an app in the same JVM to find the routers (and chose between which one they want)
2004-07-23 17:36:29 +00:00
ccb2600e67
when measuring capacity, consider data updated within the last hour as good, not just the last 5 minutes
2004-07-20 04:11:33 +00:00
f06e21ff5a
null check (oops)
2004-07-20 04:10:33 +00:00
bb0817a2ec
erg, expose the capacity calculator
...
(the last Router commit is a mod that ugha requested, but i think its ugly so its disabled atm)
2004-07-20 03:35:36 +00:00
6911f865ca
expose the capacity calculator
2004-07-20 03:34:52 +00:00
fe28b2732c
simple error condition check
2004-07-20 03:28:43 +00:00
e8e8c37496
* implement new 'capacity' concept, which replaces the old 'reliability'
...
one for peer selection and organization. reliability is kept around
for the moment and shown on the router console, but only to provide a
comparison (it is not used in any way)
* new stat in the TunnelHistory: failRate
* coallesce TunnelHistory stats (duh!)
* new ProfileOrganizer CLI ("ProfileOrganizer[ filename]*"
* implement reasonable 'failure' logic - if they are actively rejecting
tunnels or tunnels they've agreed to are failing, mark them as failing
* when choosing peers to test, exclude all fast ones
2004-07-20 03:27:34 +00:00
ef0f1ca1e7
include a lil more eye candy on the console (how active each tunnel is and last test time)
2004-07-20 02:57:55 +00:00
31ca34b954
rate.getAverageValue returns the average of the last fully completed period, but we want to include the current partial period as well
2004-07-20 02:53:41 +00:00
130399a1e7
0.3.2.3 (coming soon to a hard drive near you)
2004-07-16 21:12:27 +00:00
f0b6cbaf89
logging
2004-07-16 19:14:39 +00:00
707b173e77
differentiate between an explicit tunnel rejection (due to overload, etc) and an implicit one (the request timed out, the tunnels delivering the request failed, etc)
...
also, within the implementation of the profile, only mark the explicit rejections as a rejection
2004-07-16 00:17:28 +00:00
4381bb5026
don't rip the peer's head off after multiple tunnel rejections - penalize them *once* for the instance (not once *per* instance)
2004-07-16 00:15:34 +00:00
806d598a04
_context/getContext()
...
(missed one)
2004-07-15 05:23:18 +00:00
e737e5c950
* work around the disagreement between different versions of sun's compiler and JVM:
...
Some of them think that its ok for an inner class of a subclass to access protected data of the
outer class's parent when the parent is in another package.
Others do not.
Kaffe doesn't care (but thats because Kaffe doesn't do much for verification ;)
The JLS is aparently confusing, but it doesnt matter whether its a code or javac bug, we've got to change the code.
The simplest change would be to just make the JobImpl._context public, but I loath public data, so we make it private and add an accessor
(and change dozens of files)
whee
2004-07-15 05:12:37 +00:00
bbcde2f52b
0.3.2.2 (a lil installation testing and then i'll push)
2004-07-15 01:08:54 +00:00
f6ef77429c
some boundary cases for the queue pumper's wait time
2004-07-15 01:04:13 +00:00
71a6cf4ee6
* adjust the algorithm to deal with IO bound requests:
...
if more tokens become available while the first pending request is still blocked on
read/write (aka after allocation and before next .waitForAllocation()), give the tokens
to the next request
* refactor the satisfy{In,Out}boundRequests methods into smaller logical units
2004-07-14 21:07:57 +00:00
744ce6966f
add a new throttle (and stats) based on send processing time
...
high send processing time and low job lag means the latency is coming from outside the jobQueue - aka bandwidth throttling
2004-07-14 20:01:40 +00:00
d25cec02c2
clean up sorting for peer reliability
...
increase penalties for tunnel rejection, and keep track of the 10 minute rate as well as 1 and 60
2004-07-14 19:56:38 +00:00
f02bf37fd3
stats and stats and stats
...
track the total allocated bytes correctly (even if we're throttled)
2004-07-14 19:54:04 +00:00
304b9d41d7
on kaffe i've periodically seen some hangs in the jobqueue, so lets try being a bit more conservative with the synchroniation, and include some debugging output in the router console to help track it down (if this doesnt fix it)
2004-07-13 20:19:28 +00:00
2d6af89f60
safer operation (for use in the sim where some things aren't always availble)
2004-07-13 20:17:15 +00:00
d5ad56c4de
use smaller writes to make it look more normal
2004-07-13 20:14:18 +00:00
4f1f2cc99e
since people are using small buckets, the 10s replenish frequency is a really really bad idea (so default to 1s)
2004-07-13 05:49:16 +00:00
da439dd127
sanity checking for a kooky race condition
2004-07-12 21:33:32 +00:00
1375d01bdf
new bandwidth allocation policy and usage to include support for partial allocations (and in turn, partial write(...)) while still keeping the FIFO ordering
...
this will give a much smoother traffic pattern, as instead of waiting 6 seconds to write a 32KB message under a 6KB rate, it'll write 6KB for each of the first 5 seconds, and 2KB the next
this also allows people to have small buckets (but again, bucket sizes smaller than the rate just don't make sense)
2004-07-12 21:09:05 +00:00
7b9db07f13
target=1.3 and source=1.3, not target=1.1 and source=1.3
...
(this is what caused the runtime errors on sun jvms but not on kaffe)
((aka i slacked and didn't test sufficiently. off with my head))
this now builds and runs fine in sun 1.3-1.5 jvms, as well as kaffe
2004-07-12 16:39:22 +00:00
0f60ac5acf
0.3.2.1 (backwards compatible blah blah blah)
2004-07-11 18:57:01 +00:00
c28f19fe8a
less painful and/or redundant penalties for failures
2004-07-11 18:50:23 +00:00
3bc0e0fc8a
added source and target declarations for the javac commands so we can build with the 1.5^W5.0 JDK
...
(also added deprecation, since, well, we can :)
2004-07-11 04:16:59 +00:00
eb0e187a54
throttle tunnel participation based on whether we've had to throttle our network connection > some number of times in the last 10-20 minutes (rather than a simple "are we throttled *right now*?")
2004-07-10 22:28:34 +00:00
591dfc961e
give the reliability more positive influence so it doesn't go negative so easily
...
update the peerProfile's CLI to make the resulting stats easier to read
2004-07-10 04:15:51 +00:00
809e12b034
logging
2004-07-10 04:13:42 +00:00
4888207eca
if a client reconnects, we always want to get a new leaseSet ASAP (even if the pool hadn't been marked as stopped yet)
...
logging
2004-07-10 01:46:57 +00:00
294cb96107
if the job's startAfter is changed, tell the jobQueue to go through the timed jobs again in case the new time changes the scheduling
2004-07-10 01:44:27 +00:00
b648fa2b70
send the stats page out in chunks (more mem efficient, blah blah blah)
2004-07-10 01:39:54 +00:00
ab99122211
render status HTML in pieces (continued)
2004-07-09 05:33:19 +00:00
dd014fee88
send the router console out bit by bit rather than building it all up and sending it (thereby reducing its memory footprint dramatically)
2004-07-09 05:29:02 +00:00
c81f864de3
reduce the throttle threshold from 5s lag to 2s lag
2004-07-09 05:22:29 +00:00
90fe7dceec
include the expiration in the error message if its dropped
2004-07-09 05:20:26 +00:00
