I never noticed this, actually. The standard net.Dial(), for "historical
reasons", just prefers IPv4 instead of relying on the local resolver. This
prompted a "Wow, okay" from B. Fitzpatrick.
(https://github.com/golang/go/issues/20911).
It looks like one can get around that by using the
Fwiw I tried irtt with a similar setup, using hosts defined in /etc/hosts with
both IPv4 and IPv6 addresses:
10.72.0.230 apu2a_6
a::601 apu2a_6
10.72.0.233 apu2c_6
a::603 apu2c_6
All of the following worked as expected:
```
sysadmin@apu2a:~$ irtt client -n apu2c_6
> LGTM, but could you please squash the changes into a single commit? :)
Squashed and force pushed, so hopefully that's fine.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@heistp commented on this pull request.
> @@ -1682,7 +1683,13 @@ def check(self):
if self.ip_version is not None:
args.append("-{}".format(self.ip_version))
-args.append(self.host)
+try:
+ipaddress.IPv6Address(self.host)
+
@heistp commented on this pull request.
> @@ -1682,7 +1683,13 @@ def check(self):
if self.ip_version is not None:
args.append("-{}".format(self.ip_version))
-args.append(self.host)
+try:
+ipaddress.IPv6Address(self.host)
+
@heistp pushed 1 commit.
7d7a6ac0c5aeef4206d6d43afe735d92adeb18bc util: Add optional bracket_v6
parameter to normalise_host
--
You are receiving this because you are subscribed to this thread.
View it on GitHub:
The proposed fix is in pull request #239.
I tested an rrul test briefly and successfully passed in a link local IPv6
address containing a scope (%dev), so that should work at least as far as irtt
is concerned.
--
You are receiving this because you are subscribed to this thread.
Reply to this
Hi, I guess that creates a bit of a hassle on the flent side as it needs to
know the address type in order to surround it with brackets. It comes from Go's
requirement for brackets around IPv6 address literals in net.Dial(). I added an
issue for irtt, but I don't know if/when that will ever get
That makes sense to me, and should be supported in the gui with no changes? So
far this isn't urgent for me now so I probably won't get to it soon, so either
you can do it if you have time and it's easy, or else I'll come back to it when
needed later. :)
--
You are receiving this because you
Ok, let's see if I finally got it right.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/pull/205#issuecomment-626746340___
Flent-users mailing list
So that's there. Might have just needed an rrul_be test with socket stats, not
sure.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@heistp pushed 1 commit.
637b14c0bea66d7e3c9e3a326f2b166c7738ad33 Add all_scaled_delivery to
PLOTS_MAY_FAIL and a sample data file.
--
You are receiving this because you are subscribed to this thread.
View it on GitHub:
In a plot that has subplots, like all_scaled, it appears that --bounds-y can
only be specified once, and it applies to all of the subplots for the plot. Is
there a way to specify bounds separately for each subplot?
--
You are receiving this because you are subscribed to this thread.
Reply to
@heistp pushed 1 commit.
ebf8aa72bc1edd8e69663afebee96e98f3462878 Update delivery rate title for
all_scaled_delivery.
--
You are receiving this because you are subscribed to this thread.
View it on GitHub:
Hi, just adding a few plots were using with rrul_be...
Signed-off-by: Pete Heist p...@heistp.net
You can view, comment on, or merge this pull request online at:
https://github.com/tohojo/flent/pull/205
-- Commit Summary --
* Add TCP delivery rate plots (tcp_delivery_rate
Thanks! Works for me, only I got a little confused because 'tcp_cong_control'
stopped working and is replaced by 'cc_algo'. It's no problem for me as I can
adjust and it's more consistent now, just that people's existing batches may
behave differently...
--
You are receiving this because you
I can set 'markings' for tcp_nup, but it seems that setting 'marking' on a
tcp_1up test is not working. Any chance you could take a look or point me where
to fix this?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Should be fixed. I admit that I haven't actually tested with a cutoff, but the
plot works...
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@heistp pushed 1 commit.
f736e627d6a6019aca6253d83221d6fb5d0b0d0e tests: Add cutoff for upload_box plot.
--
You are receiving this because you are subscribed to this thread.
View it on GitHub:
Signed-off-by: Pete Heist p...@heistp.net
You can view, comment on, or merge this pull request online at:
https://github.com/tohojo/flent/pull/200
-- Commit Summary --
* tests: Add tcp_delivery_with_rtt and upload_box plots.
-- File Changes --
M flent/tests/tcp_1up.conf (15)
M
So as it turns out, it would still be useful in this case to be able to plot
the box plots for each flow in a tcp_nup test. Is there a way to do that?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Closed #198.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/issues/198#event-3042804193___
Flent-users mailing list
Flent-users@flent.org
Ah, I understand now that that the boxes would appear when there are multiple
repetitions in the batch run. I thought they might show the variances in data
for each flow. While that could be useful, it looks like I can get what I
actually need at the moment by adding other open data files in
In flent from head (1.9.9-git-3d663cf) I seem to be missing the boxes on box
plots, like the attached image. I'm mostly doing tcp_nup and tcp_1up. It's the
same on Linux and MacOS. I'll send a .flent.gz in case that helps.
![Screen Shot 2020-02-16 at 11 27 22
Sent a pull request but only for setting it globally for TCP so far, as setting
it per-flow and for the UDP packet size was more than I bargained for. :)
The reason I'm probably seeing this more than most people is because I have
non-default settings for `net.ipv4.tcp_wmem` for doing tests on
This is only a start on what we discussed, as its over my head at the
moment how to make this setting per-flow and have it affect the UDP packet
size. Sorry if short on time at the moment to dig into it further. :)
Signed-off-by: Pete Heist p...@heistp.net
You can view, comment on, or merge
Ah ok, so we could combine them for now and there can always be separate flags
later if needed, plus that's good if the UDP one can be overridden with the
test param. Will see if I can make this change...
--
You are receiving this because you are subscribed to this thread.
Reply to this email
Ok, it sounds good allowing the comma separated values for setting it per-flow,
but I'm not sure about sharing it with the UDP setting. People usually specify
a UDP packet size that's something below the MTU right? That would cause pretty
high CPU on the stream test. I expect when people are
Sometimes it can happen that you see periodic dips in throughput plots, like
this:
![netperf_send_size](https://user-images.githubusercontent.com/4397233/73930197-7157c480-48d6-11ea-8423-c7c90a48cfb3.png)
The period of the dips can be short, depending on test conditions / parameters:
Closed #179.
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/issues/179#event-2617445990___
Flent-users mailing list
Flent-users@flent.org
Added in 2d0d8fbd844b2f2153f098e157294983e3432dc7, thanks for accepting.
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/issues/179#issuecomment-529311035___
Re-arranged gen_filename a bit now that some parts of the filename are optional.
Signed-off-by: Pete Heist p...@heistp.net
You can view, comment on, or merge this pull request online at:
https://github.com/tohojo/flent/pull/180
-- Commit Summary --
* batch: Add --batch-no-timestamp command
Currently, the output filenames for batches always include `batch_time`:
```
def gen_filename(self, settings, batch, argset, rep):
filename = "batch-%s-%s-%s" % (
settings.BATCH_NAME,
batch['batch_time'],
batch.get('filename_extra', "%s-%s" %
Adding two plots that Im frequently using.
Signed-off-by: Pete Heist p...@heistp.net
You can view, comment on, or merge this pull request online at:
https://github.com/tohojo/flent/pull/176
-- Commit Summary --
* Add totals_with_tcp_rtt and upload_with_ping_and_tcp_rtt plots.
-- File
No other reason for the order except it felt natural to me that what comes
later overrides what comes earlier. It's about like which way do you like your
mouse wheel to scroll ("finger on the paper" is correct, obviously :).
I'm totally fine with taking out the `reversed` call for priority
. :)
Signed-off-by: Pete Heist p...@heistp.net
You can view, comment on, or merge this pull request online at:
https://github.com/tohojo/flent/pull/175
-- Commit Summary --
* Add multiple inheritance support for batch files.
-- File Changes --
M flent/batch.py (11)
-- Patch Links --
https
You can view, comment on, or merge this pull request online at:
https://github.com/tohojo/flent/pull/174
-- Commit Summary --
* metadata: Add pacing sysctls (ca_ratio and ss_ratio)
-- File Changes --
M flent/metadata.py (2)
-- Patch Links --
Jon and I noticed this with pacing and delivery rate, and this plot might be
useful for others as well. :)
You can view, comment on, or merge this pull request online at:
https://github.com/tohojo/flent/pull/168
-- Commit Summary --
* Fix calculation of pacing and delivery rate.
* Add
```
export QT_API=pyside2
```
A space had crept in my previous post after the equals. :)
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
It looks like it was still using PyQt5 in your case, so make sure to first do:
```
export QT_API= pyside2
```
If this is a long-term fix for Mac, maybe that var could be set on MacOS when
sys.platform is darwin, assuming we have a chance to do so before those libs
are initialized. Probably
Ah, ok, well it looks like PySide2 has a more permissive license anyway,
although I guess that shouldn't affect flent. I haven't noticed any other
problems using it, although my testing has hardly been extensive.
--
You are receiving this because you commented.
Reply to this email directly or
Ok, the problem doesn't happen with PySide2. It's likely with recent versions
of PyQt5, although I'm not exactly sure of the first version where it started.
It's also possible the bug doesn't happen on MacOS 10.14 (Mojave), but my 2011
hardware is pinned to 10.13 so I can't test that. So far
Yes, it does, then the symptom after the error (same with one or more files) is
that Python is unresponsive and there's 100% CPU usage.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Against better judgment I upgraded PyQt5 to 5.13.0 on both the Mac and Linux
and still see the problem on Mac but not on Linux, so that suggests it's not
directly related to the PyQt 5 version.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or
Hmm, that made me realize it's working in Linux but not on my Mac, with
slightly different versions of Python and PyQt5:
Not working (on MacOS 10.13.6):
```
% flent --version
Flent v1.9.9-git-83d14d6.
Running on Python 3.7.4 (default, Jul 9 2019, 18:15:00) [Clang 10.0.0
(clang-1000.11.45.5)].
I compiled on some new test boxes from head (83d14d68) and ran a test. flent -i
for the text output worked but when I use flent -gui I see the following errors
before it freezes:
```
ERROR: Error while loading data file: '
returned a result with an error set'. Skipping.
ERROR: Error while
On flent from head I'm not able to generate summary text for a tcp_2up test:
```
% flent -i *.flent.gz
Started Flent 1.2.2-git-325dad7 using Python 3.7.3.
Summary of tcp_2up test run from 2019-07-08 19:53:28.768995
Title: 'sce cc:reno,reno-sce q:triple-isolate bw:50Mbit rtt:80ms crtt:100ms'
Super, works for me.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/issues/165#issuecomment-508557242___
Flent-users mailing list
Yes, that's the case, so it's working now with the kernel compiled with
CONFIG_INET_DIAG and as root. Closing this, unless you want me to test anything
else...
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Ok, there is a difference there (below, sudo version contains pid at end of
ESTAB line).
Running flent with sudo works even though I didn't used to need to do that on
older Ubuntus. It's a fine workaround for now, only in addition, ss with my
custom kernel doesn't show TCP RTT or a lot of
I have ss installed and am using --socket-stats, but for some reason tcp_rtt
plots don't seem to be working (using rrul_be but I need it ultimately for
tcp_1up_noping). I'm using flent from HEAD and python 3.7.3 on Ubuntu 19.04. Do
the attached .flent.gz files and output from --verbose show
> On Jan 11, 2019, at 5:31 PM, Pete Heist wrote:
>
>> On Jan 11, 2019, at 5:00 PM, Dave Taht wrote:
>>
>> I just turned it on again (I think it had crashed).
I actually never noticed that it went on, and it isn’t on now. Is it really
When I make a tcp_rtt_box_combine plot of ten data sets, it looks like this:
![tcp_rtt_box_combine](https://user-images.githubusercontent.com/4397233/50589846-8dde2d80-0e88-11e9-90c6-85664b8d4701.png)
but what I make a tcp_rtt_bar_combine, it looks like this:
Ah, I see TSDE is Pollere's work. I need to go through the talks referenced on
pollere.net asap to get smarter on that. Will be on some roofs today though,
p2p connection for the neighbors...
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
As for packet loss and reorders, there's the `lost` property on each
`round_trip` that could be plotted, but for re-orders there's so far just a
global `late_packets`, which is the number of packets who sequence number is
lower than the previous one received. It would be possible to add a
Ok, well if we do go for it, so far in irtt's JSON there's just an average
`send_rate` and `receive_rate` under stats, both of which contain an integer
`bps` and a `string` text representation. `send_rate` ignores lost packets and
`receive_rate` takes them into account. Let me know if anything
> On Aug 27, 2018, at 1:40 PM, flent-users wrote:
>
> Hi Pete,
>
> > On Aug 25, 2018, at 19:53, Pete Heist wrote:
> >
> > 50ms wouldn't be too disruptive in most cases. At 1 Mbit, the 5% of
> > bandwidth threshold is crossed.
>
> Would it be possi
Yeah, I haven't thought about what it would mean to change the semantics of the
existing tests by changing the default interval. Although, as it is, there's
still a fallback to UDP_RR for current tests, so results can change if irtt
isn't installed or the server isn't reachable for some reason.
> On Aug 25, 2018, at 5:43 PM, Dave Taht wrote:
>
> " Flent has been an invaluable network analysis tool while performing
> these experiments and has become my goto application for testing
> networking performance after making changes/optimizations/tweaks etc.
>
> from -
Is there an argument for lowering the default interval for when flent calls
irtt?
The default irtt packet length with no payload is 60 bytes, so here are
bitrates at various intervals for IPv4+Ethernet (106 byte frames, and tripled
for RRUL's three UDP flows):
```
200ms => 4.2 Kbit * 3 = 12.7
> On Jun 26, 2018, at 12:43 PM, Toke Høiland-Jørgensen wrote:
>
> There's a limit to the maximum aggregation size on VO and VI (for VO it
> is '1 packet' as Sebastian noted). Minstrel will take this into account
> when building the aggregate, but IIRC only for the first entry in the
> retry
> On Jun 26, 2018, at 10:40 AM, Sebastian Moeller wrote:
>
> Hi Pete,
>
>> On Jun 26, 2018, at 00:40, Pete Heist wrote:
>>
>> The rrul test over a point-to-point WiFi link cuts the total TCP throughput
>> considerably below that of rrul_be. For example,
The rrul test over a point-to-point WiFi link cuts the total TCP throughput
considerably below that of rrul_be. For example, on an 802.11n 20MHz MCS 15
link:
rrul_be: ~90mbit
rrul: ~30-45mbit
I think this came up before, but could the standard rrul test be exceeding the
802.11e spec in terms
Currently the voip-rrul test is actually an rrul_be + voip test, but it could
be useful to sometimes do a voip-rrul test that’s rrul (not best-effort only) +
voip.
Would it make sense to make the current voip-rrul use the rrul test instead,
and to add a voip-rrulbe test for the best-effort
> On Jun 10, 2018, at 5:32 PM, Toke Høiland-Jørgensen wrote:
>
>> I had good luck with reCAPTCHA solving spam for one simple form. Every
>> time I click “I am not a robot" I fear it’s somehow snapshotting my
>> soul, but that’s another thing… :)
>
> Well as someone that runs uMatrix and
> On Jun 10, 2018, at 4:37 PM, Toke Høiland-Jørgensen wrote:
>
>> Anyway…I don’t know mailman from the admin side, but what about trying
>> to configure reCAPTCHA support? It looks like they added it to mailman
>> 2.1.26 (setup instructions in release notes):
>
> Ah, that might be an option;
> On Jun 10, 2018, at 3:42 PM, Toke Høiland-Jørgensen wrote:
>
> Just a note that I have been forced to delist the Flent mailing lists
> from the public mailman listing page. Apparently spammers have started
> using unsolicited mailman confirmation requests as a way to spam people;
> so I had
> On May 6, 2018, at 9:17 PM, Sebastian Moeller wrote:
>
> you could also try different packet sizes for you probes. The size ratio
> should give you an idea about the RTT difference explained by simple transit
> times and anything on top of that might be related to queueing.
> On May 6, 2018, at 10:51 PM, Pete Heist <p...@eventide.io> wrote:
>>
>> Ah, so this is a wired connection? And you are only targeting a
>> particular setup?
>
> The backhaul is a mixture of wired and wireless links, each with an ALIX or
> APU router.
> On May 6, 2018, at 9:27 PM, Toke Høiland-Jørgensen wrote:
>>
>> The backhaul I’d like to test it on uses mostly NSM5s as the wireless
>> devices and APUs for routing, QoS, etc. The QoS scripts use the htb,
>> sfq and prio qdiscs. I’m hoping I can just add a prio qdisc / tc
>>
> On May 6, 2018, at 5:54 PM, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>
> Pete Heist <p...@eventide.io> writes:
>
>> Channel contention delay may be estimated by the difference between
>> the round-trip times for the strict priority VO and BE packets (0
> On Feb 19, 2018, at 11:36 AM, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>
> Pete Heist <petehe...@gmail.com> writes:
>
>> I liked the simplicity of RFC 795 (https://tools.ietf.org/html/rfc795) from
>> 1981 :) so, mostly in jest, I’ll propose a new syste
> On Feb 18, 2018, at 9:03 PM, Sebastian Moeller wrote:
>
> Really, I am always confused by this especially since I aways forget
> whether network order is big or little endian (looked it up again it is
> big-endian).
Here I got nailed by assuming that ECN being in
> On Feb 18, 2018, at 7:11 PM, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>
> Pete Heist <petehe...@gmail.com> writes:
>
>> The author pulled the probe in pretty quickly, so it’s already upstream.
>> Some samples:
>>
>> https://www.drh
<https://www.drhleny.cz/bufferbloat/smokeping/irtt_smokeping2.png>
Maybe I’ll post another image after a while if something more exciting happens…
:)
> On Feb 18, 2018, at 12:29 AM, Pete Heist <petehe...@gmail.com> wrote:
>
> I wrote a SmokePing (https://oss.oetiker.ch/smokeping/
> On Feb 18, 2018, at 10:55 AM, Sebastian Moeller wrote:
>
> Not that I have looked at the actual parameter names, but how about having
> --dscp mask out the upper two bits (if they are set) and --tos allowing to
> set the full legacy 8 bit?
I think that’s the right thing to
> On Feb 18, 2018, at 10:38 AM, Toke Høiland-Jørgensen wrote:
>
> Yes, that would probably be a good idea; and also agreement on the values
> would be a good idea ;) the current mapping I put in is taken from netperf,
> which in turn cites an RFC. But it's different from the
> On Feb 17, 2018, at 11:21 PM, Toke Høiland-Jørgensen wrote:
>
> Sebastian Moeller writes:
>
>> Hi Toke,
>>
>> I made the mistake of updating both my os (from macos 10.12.6 to
>> 10.13.3) and flent (to 1.2 I believe) at the same time. Now when I try
>> to
I wrote a SmokePing (https://oss.oetiker.ch/smokeping/) probe for IRTT:
Code:
https://github.com/peteheist/SmokePing/blob/master/lib/Smokeping/probes/IRTT.pm
Pull request: https://github.com/oetiker/SmokePing/pull/110
I hope to use this in my ISP, which uses SmokePing for about 70 access
> On Feb 9, 2018, at 11:36 AM, Toke Høiland-Jørgensen wrote:
>
> Sebastian Moeller writes:
>
>> Hi Toke,
>>
>> Ah, great, this got me a bit further:
>>
>> UDP RTT test: Cannot use irtt runner (Irtt connection check failed:
>> [InvalidServerRestriction] server
>
> On Feb 8, 2018, at 8:53 PM, Sebastian Moeller <moell...@gmx.de> wrote:
>
> Hi Pete,
>> On Feb 8, 2018, at 20:49, Pete Heist <petehe...@gmail.com> wrote:
>>
>> It should work, but the irtt server needs to be running. On your box where
>> n
It should work, but the irtt server needs to be running. On your box where
netserver is running (sounds like the Mac for you if you’re testing locally),
run “irtt server”.
On Linux you can use the irtt.service file in the repo with systemd to run it
at startup. I’m working on a Debian package
> On Jan 30, 2018, at 9:17 AM, Dave Taht wrote:
>
> https://github.com/lucas-clemente/quic-go
Cool, in case it ends up in flent one day I’d like to add that to the “flenter”
tests.
Two things I noticed:
- It’s possible to use BBR with QUIC (not implemented in quic-go
Cool, hope that works...
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/pull/125#issuecomment-351200608___
Flent-users mailing list
For some reason, the Go pflag packages (both of them) require long form flags
with values to have equals signs. This works for me, but I'm not sure of and
didn't test the local_bind stuff in particular, so by all means fix or replace
any of it if there's a better way. :)
You can view, comment
If anyone would like to try a semi-public irtt server for testing, I’ll leave
it up for a bit:
irtt client -hmac flentusers irtt.eventide.io
irtt can be had here, where you can also report any problems:
https://github.com/peteheist/irtt
The server is configured with the hmac key above, and
Yep, that fixed it (three runs an no warning), thanks!
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/issues/124#issuecomment-346933882___
By the way, in case you'd want undocumented flags for both client and server
for dropping or duplicating some percentage of packets I can add that. I
currently have compiled in constants for those.
--
You are receiving this because you are subscribed to this thread.
Reply to this email
I'm filing this here for the moment but we can move it to irtt if it turns out
to be something for me. The warning says:
Unable to get packet timestamp and seqno: u'wall'
The seqno does some to be there though.
It seems to happen when `lost` is `true_down`, which would mean there aren't
any
Closed #123.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/issues/123#event-1357443251___
Flent-users mailing list
Flent-users@flent.org
Duh, I didn't have an up to date irtt in my test bed. `-v` was helpful, thanks.
Maybe I need to set up my test bed to update irtt on commit.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Actually not sure if this should work yet or not, but rrul_be_nflows is one of
the more common tests I do and it still uses UDP_RR...
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Oh yeah, probably time for this issue thread to retire. :)
So I'm glad! Looking forward to playing with this more soon. Thanks for all
that refactoring too, looks like it was some real walking through walls...
--
You are receiving this because you commented.
Reply to this email directly or
> On Nov 22, 2017, at 8:49 AM, Toke Høiland-Jørgensen
> <notificati...@github.com> wrote:
>
> Pete Heist <notificati...@github.com> writes:
>
> >> > And this likely takes the mean value of all transactions and
> >> > summarizes it at t
> On Nov 21, 2017, at 10:56 PM, Toke Høiland-Jørgensen
> <notificati...@github.com> wrote:
>
> Pete Heist <notificati...@github.com> writes:
>
> > Trying to confirm how latency was being calculated before with the
> > UDP_RR test. Looking at its
Trying to confirm how latency was being calculated before with the UDP_RR test.
Looking at its raw output, I see that transactions per second is probably used
to calculate RTT, with interim results like:
```
NETPERF_INTERIM_RESULT[0]=3033.41
NETPERF_UNITS[0]=Trans/s
NETPERF_INTERVAL[0]=0.200
> On Nov 21, 2017, at 3:53 PM, Toke Høiland-Jørgensen
> wrote:
>
> > Next thing I noticed as for current tests, for rrul_be_nflows, the
> > test completed but only one irtt instance ran (also just saw one
> > connection to the server).
> >
> > % flent rrul_be_nflows
> On Nov 21, 2017, at 11:36 AM, Toke Høiland-Jørgensen
> wrote:
>
> Ha! Epic fail! :D
>
> Well, I only just managed to finish writing the code and unbreaking the
> CI tests; didn't actually get around to running any tests. I've fixed
> those two errors, and am
> On Nov 20, 2017, at 10:14 PM, flent-users wrote:
>
> Winstein plot of latency variance? It doesn't get denser, it gets darker.
>
> Packet loss vs throughput?
Not sure what that is exactly. Something like from July 2014 on this page?
> On Nov 20, 2017, at 9:58 PM, Toke Høiland-Jørgensen
> wrote:
> Okay, testable code in the runner-refactor branch.
>
> Ended up doing a fairly involved refactoring of how runners work with
> data; which is good, as the new way to structure things makes a lot more
>
1 - 100 of 123 matches
Mail list logo