Options to enable and disable tcpdump and tcp_logger
Committed in https://sourceforge.net/p/teacup/code/ci/51a44bf28a0763a24b4b2d97e415564530e02aae/
New config.py options:
Attached is an updated patch -- using TPCONF_tcpstats_logger (instead of TPCONF_tcp_logger) as a config variable to enable/disable TCP-related stats logging.
Options to enable and disable tcpdump and tcp_logger
Options to enable and disable tcpdump and tcp_logger
Post-release, merged v1.1 back into default branch
Added tag v1.1 for changeset e1d110a7fd4e
Updated changelog
Tweak make_tar.sh to build from correct repo branch
Add doc pdfs locally, update README, AUTHORS and INSTALL
Add jtkua to AUTHORS
Update the release tarball creation script
Update TODO
Update Changelog and VERSION
Ack Isaac, Djuro and Russell wrt teaplot development and tweaks
Remove unused pktloss.py
fix id tag in pktloss.py
Small editorial changes in docs
Embedded fonts in all figures/graphs of the docs
Update ChangeLog
tidy up (deletion) of docs/teaplot-docs.tex
Clean (removed change tracking) versions of updated documentation
Bug fix handling of noclients_game in start_fps_game()
Substantial updates to documentation for cover key changes since 1.0 release
Fix typo in ttprobe handling
New {extract,analyse}_owd and new {extract,analyse}_pktloss functionality, some small code cleanups
Updated ChangeLog
mention pktgen-0.3.1 in Changelog
Update Changelog to reflect ticket#14
experiment runs for longer than needed
Applied in https://sourceforge.net/p/teacup/code/ci/75624d28a630951ef24056da0f07a64bf05455d1/
Ticket#14, improve how experiment run time is calculated
Plotting time serie data points of the first flow twice
I have updated all the plot() code, but boxplot() is unchanged as it is not clear from the R man page whether panel.first works as expected (need to test) and for boxplot() plotting the first series twice is much less of an issue due to the relatively small size (assuming not too many outliers).
Optimised plotting by avoiding the double plotting of the first series
experiment runs for longer than needed
Ok. Sounds like a good idea. The code also looks good to me. Have you updated the documentation? I assign this to Grenville as you guys are closer to each other :).
That's true. However, I included the experiment duration in the calculation of the total duration just in case the user intentionally wants to run the experiment for extra time (e.g. to launch external traffic generators for testing purposes). If you see that this is not th case, I can remove the duration from total duration calculation.
I appreciate that somebody finally cleans this up :). But shouldn't the duration simply be max(TrafficGen_wait_time[i] +TrafficGen_duration[i]) for i = 1 to number of traffic generators? Otherwise, we would still potentially have extra wait time, e..g if somebody defines duration but then does use a shorter constant duration for the first and only defined traffic generator.
ymin argument of boxplot has not effect
ymin argument of boxplot has not effect
Patch aplied, https://sourceforge.net/p/teacup/code/ci/392986c7a556e18551d21f247de3d1da8192cde5/
Fixed setting ymin for boxplots, ticket#12
Thanks, that is a good optimisation which is long overdue. We should apply this to all plot() though. I will work on it.
Plotting time serie data points of the first flow twice
Looks good to me, please apply.
experiment runs for longer than needed
Plotting time serie data points of the first flow twice
ymin argument of boxplot has not effect
add missing nginx config templates
Add support for launching nginx server
Add support for launching an nginx server (jtkua@swin.edu.au)
remove existing files before running experiments
TCP TIME_WAIT conditions causing nttcp runs to sometime not occur
Sorry for the delay, I've belatedly applied this patch in https://sourceforge.net/p/teacup/code/ci/8e39f67869763b6d6d32109d934df10994a75d82/
Apply TIMEWAIT adjustments from Ticket#10
TCP TIME_WAIT conditions causing nttcp runs to sometime not occur
Added .hgeol to normalise .lyx line endings whe...
Updated teaplot (fab animate) to version 0.2 wi...
Add support for generating DASH traffic with da...
Patch applied at https://sourceforge.net/p/teacup/code/ci/3b086222a74b5c882cd5ce...
Adding FreeBSD AQM support
Commit support for AQM in FreeBSD bottleneck, p...
Adding FreeBSD AQM support
remove existing files before running experiments
Committed, https://sourceforge.net/p/teacup/code/ci/9b2a6ac01b080d57b6b20b1e5bb2...
clean-out any locally pre-existing files for gi...
I'll push this commit soon
That's correct. Here is the updated patch: diff -Naur a/experiment.py b/experiment.py...
I think we don't need the "if len(glob.glob(file_pattern)) != 0:" check right?
Oh yes, apologies, must have copied and pasted incorrectly. Here is the corrected...
The os.remove(f) seems insufficiently indented, but otherwise okay by me. cheers,...
remove existing files before running experiments
Small tweaks to analysis code and comments rela...
Make extract_pktloss tidier. Remove intermediat...
typo
Add kernel code for ttprobe (event-driven Linux...
Somewhat belatedly, we've now committed pktloss.py and some small changes to analyse.py...
Somewhat belatedly, we've now committed pktloss.py and some small changes to analyse.py...
Add tools/pktloss.py and fix analyse.py so extr...
checking curr stays valid to host_times in clockoffset.py
checking curr stays valid to host_times in clockoffset.py
Done
Trying a fix to adjust_timestamps
This fix looks good to me. You can close the ticket after this is checked in. Cheers,...
tcpdump not in path on Linux
Update trafficgens.py to use new pktgen functio...
Add pktgen-0.3.1.tgz from BITSS project, approv...
checking curr stays valid to host_times in clockoffset.py
Updated environment variables with TC_ prefix i...