jcarr/networkQuality, branch mainUnnamed repository; edit this file 'description' to name the repository.
http://cgit.wit.com/jcarr/networkQuality/atom?h=main2024-11-16T06:08:14Zfix build2024-11-16T06:08:14ZJeff Carrjcarr@wit.com2024-11-16T06:08:14Zurn:sha1:11bbac055c2c02b474288547f0d7cd0adad1f7a2??? don't merge this2024-11-06T19:43:49ZJeff Carrjcarr@wit.com2024-11-06T19:43:49Zurn:sha1:471238a26d4607da7be7732d551911556b9df0bb
Signed-off-by: Jeff Carr <jcarr@wit.com>
build for debian2024-11-06T19:42:55ZJeff Carrjcarr@wit.com2024-11-06T19:42:55Zurn:sha1:047c382537b76df4ac582230b72c98ef5aa6cbc8todo: find a way to store these in the git commit2024-11-06T18:41:51ZJeff Carrjcarr@wit.com2024-11-06T18:41:51Zurn:sha1:5e5a66739b966272cb4bac6f9ebe72b2365f840e
Signed-off-by: Jeff Carr <jcarr@wit.com>
[Feature] Be more explicit about which part of the test was unstable2024-01-29T15:15:24ZWill Hawkinshawkinsw@obs.cr2024-01-29T15:15:24Zurn:sha1:343954285090861f84a26e9818e2e946f87b8d1f
When a test did not run to stability, tell the user which part of the
test was unstable.
Signed-off-by: Will Hawkins <hawkinsw@obs.cr>
[Feature] Special RPM calculation semantics when test does not stabilize2024-01-29T14:50:19ZWill Hawkinshawkinsw@obs.cr2024-01-29T14:49:51Zurn:sha1:60b273e673b7af9b30bf439bb1a177ee339883ba
When the test does not stabilize (responsiveness), using only the most
recent MAD probes to calculate the provisional final RPM could leave us
with a very small number of samples. So, instead, use all the probe
measurements to do the calculation.
Signed-off-by: Will Hawkins <hawkinsw@obs.cr>
[Bugfix] Do not advance the interval when stable2024-01-28T03:04:12ZWill Hawkinshawkinsw@obs.cr2024-01-28T03:04:12Zurn:sha1:59b565b068c007e0e99bfc382f32fda3bc0cfe6f
By advancing the stability algorithm's interval when the measurement is
complete, we incorrectly expelled one of those intervals from the RPM
calculation. As a result, the final RPM calculation did not include all
the probe measurements that factored into stability calculation.
Signed-off-by: Will Hawkins <hawkinsw@obs.cr>
[Bugfix] Stabilizer was optimistic2024-01-28T03:02:49ZWill Hawkinshawkinsw@obs.cr2024-01-28T03:02:49Zurn:sha1:e461216c7c560c270b1759f28adc62403f3b28e2
The "all complete" flag incorrectly overwrote a previous value, even if
that value was false. The result? If an aggregate was incomplete but any
later aggregate was complete, it would conclude that all aggregates were
complete.
Signed-off-by: Will Hawkins <hawkinsw@obs.cr>
[Bugfix] Only probes collected during stable MAD count2024-01-27T01:46:05ZWill Hawkinshawkinsw@obs.cr2024-01-27T01:46:05Zurn:sha1:e60066f572464da3a6f55999a48b1d38cf6423d6
For calculating the final RPM, only those probes that are sent/received
during the stable MAD should count.
Signed-off-by: Will Hawkins <hawkinsw@obs.cr>
[Bugfix] Continue Adding LGCs During Responsiveness Stabilization2024-01-20T00:40:11ZWill Hawkinshawkinsw@obs.cr2024-01-20T00:37:11Zurn:sha1:65b039e33717ee43620363704cb3daa304a5e724
The client was not continuing to add load-generating connections at each
stability-check interval (as called for by the spec).
Also, the client now properly obeys the `--rpm.id` option given on the
command line.
Signed-off-by: Will Hawkins <hawkinsw@obs.cr>