Age | Commit message (Collapse) | Author | Files | Lines |
|
Results if NDR search is in TPS (so unidir) since
https://gerrit.fd.io/r/c/csit/+/28208
The bug was causing all reconf tests report zero loss,
as they used half of NDR load during reconfiguration.
+ Small improvements to descriptions of other traffic keywords.
Change-Id: Ib9af5861cc09e698eb762feb3f0a019571d17962
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Jira: CSIT-1768
Change-Id: I888ae1a5754fa07297d4cdf65c2be0e3e49d89a5
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
Modify initialize L2BD ACL keywords to be usable for both 2-node and
3-node topologies and update testsuites accordingly.
Add the missing macip T-rex profile.
Add classifier tests to 2n-tx2 job specs.
Change-Id: I17b84b8fc18ef9a6f275ae0238a0665ac2017f01
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech>
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: Ib9dfb8623757fd23fd8b49b5e04ed86e202ada66
|
|
Change-Id: Iebd62fd6b0c798f7b4dd1f3b093c156e533b3900
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
Change-Id: I23541e1c07a42b4baa560c343ffeed3b963fe694
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
Signed-off-by: Tomas Alexy <tomas.alexy@pantheon.tech>
Change-Id: Ifb97088a26f3c2ab2ecee28ed43e6e957a844c64
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: Ibafaaca4438587284d1e255c764f0701002941e8
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: I7d83ba048e0609d6b8623fab5c2960e48a37c023
|
|
Add testbed file and make code changes in common.sh.
Change-Id: I493bb92db3c66358afa81ec81bcd99b296f49c03
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech>
|
|
Do not expand argument expressions into actual verbose values.
Change-Id: I3d86cd826e9cd03cefefe2e197250f76414ac101
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Jira: CSIT-1597
Change-Id: I77b746584851331d46eca1caafd4468d18f1e78d
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: Ieff60a44d42d66acee8ba1680e7e285d6cd01bc9
|
|
Disconnect+connect cycle is expensive and slow.
This change tracks connected client instances
so later "connect" to the same target uses it.
Explicit disconnects are allowed (and executed before VPP ends),
but once again disconnected instances are cached and reused,
as creating a new instance is more expensive than just connect.
+ Add missing checks on interfaces being up to appropriate keyword.
+ Use appropriate keywords.
+ Add a comment explaining why a simpler keyword is not appropriate.
+ Improve VPP checks in containers.
+ Fix the vppctl check to actually work.
+ Add PAPI check to ensure VPP is really ready.
+ Delay/reorder checks to make them faster with multiple containers.
+ Leave some TODOs to improve various lifecycles later.
+ As we do not stop VPP in test/suite teardown:
+ One final disconnect is needed, added to __init__.robot teardowns.
- Import of the final disconnect keyword is ugly, but it works.
- We could use a hashable class for distinguishing node+socket pairs.
- Are we connecting to VPP inside VMs?
Change-Id: I49cd726740c3e8cae1591c7c84b85a447241228f
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
+ Reduce time overhead when parsing --include vs --test
+ Input files will remain the same
+ 3n-hsw 150include ~24min, 150test ~5min
+ 2n-clx 489include ~61min, 489test ~9min
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: I7e5957c32385aecbc4551bd124da41afa5bc63c9
|
|
Jira: CSIT-1597
Change-Id: Ice342d6cb7cd244d712c42a3d468c851c79739f4
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
Jira: CSIT-1597
+ add ipsec_sa tear down action
Change-Id: I4d1e6f26c14e61e8ddc6f29fbf5528a034c84eb0
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
A common mistage is to add a file into GPL directory,
with Apache license only.
This adds a tox checker to vote -1 on such changes,
gpl_license.log contains the list of affected files.
There are TODOs related to unlicensed files
or unconventional GPL licenses.
Change-Id: I267e4fb98afd199d5acc7d211068b28c97b31dc9
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Jira: CSIT-1769
Change-Id: I98510cd8e627d7347f77d0032b7bac28f2e36c61
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
This reverts commit a9f54ca5080aeef17686f300a6807bf9b46b7c90.
Reason for revert: DO NOT MERGE BROKEN TESTS!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
https://logs.fd.io/production/vex-yul-rot-jenkins-1/csit-vpp-device-master-ubuntu1804-1n-skx/9918/archives/log.html.gz
Change-Id: I2ce2970a43e5dd7487eeb54d1ccbb149e06cf8fa
Signed-off-by: pmikus <pmikus@cisco.com>
|
|
Jira: CSIT-1769
Change-Id: I0ab906966f760b809b8431530c010a8f0835d62c
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
In Gerrit 30041 that was not tested properly.
Change-Id: I42f57f1933b0e7d029030f4ec502f7268780827b
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Attempt to unbind a driver from a device only if it is bound to a
driver.
Remove the dynamic addition of an existing device ID to a driver. From
the docs [0]:
Writing a device ID to this file will attempt to
dynamically add a new device ID to a PCI device driver.
Since we assume the VFs are bound to the kernel driver when VPP Device
topology creation is done, it implies that the kernel driver supports
the device ID of those VFs, removing the need to add the support.
[0]: https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-bus-pci
Change-Id: I20f3ca071a5a84a06ff358ba514532248a8f9ad0
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech>
|
|
This is a follow-up to https://gerrit.fd.io/r/c/csit/+/20394
https://gerrit.fd.io/r/20119
has changed the way archival works,
everything should now go to logs.fd.io
(instead of appearing on run page in jenkins.fd.io).
The glob pattern for archiving is quite eager,
doing recursive search. That is good, as it can find
also misplaced useful outputs.
But it also means our usage of copy_archives function
creates two copies of archived directtories,
usually archives/ and archives/archive/.
This change renames copy_archives to move_archives,
with few workarounds to support multiple calls.
I also renamed ARCHIVE_DIR value from $CSIT_DIR/archive
to $CSIT_DIR/archives to make "move" operation look natural.
Finally, download_builds function is being removed,
as after recent improvements to VPP compilation speed
nobody seems to be using it.
Change-Id: I19a429e1dfdfaab7fcf32a9609963b1aebd33c6c
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Jira: CSIT-1597
Change-Id: Ib48d20ba087bc9828970c3acc081ac6e4447c20c
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
+ Move "|| true" to a place that really aviods errors.
+ Attemp to parse decoded string if trigger is not found in plain one.
Change-Id: If3587229ec588f9ad41acb3050add1142032d2d8
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Ticket: CSIT-1771 VPP-1207 VPP-1675
Change-Id: I8ba2d62054361e72b833943327434ea071e7e568
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
+ The old command is tried if the new one fails.
Change-Id: I0afdcd7ac699595347b42f3a7d34e1281bb0dc0e
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Even with IPv6 nd disabled in VMs, there are IPv6 packets coming from
DuT:
--dut_if2_mac ba:dc:0f:fe:02:05
...
dst = 33:33:00:00:00:01
src = ba:dc:0f:fe:02:05
Docker disables IPv6 by default on interfaces it manages, which doesn't
include VF interfaces. Disabling IPv6 in containers using sysctl may fix
this.
Change-Id: Idba506653442a3da2cd82e69cedddb3294d92788
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech>
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: I46750f8571d86dc13df95252e8894c4b988efd5c
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: Id56b87ab868f2897a6563914b0beca2acc25e706
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: I145a4b5511141f1e2b4e387daa358e32dd2c8015
|
|
The previous code counted full TCP connections,
which need one more packet, leading to worse results.
Change-Id: Ifcf78356b6ed54819ea0bf5aa069d7d9cb951183
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
16 times faster setting at cost of 50% larger archive.
Before:
$ time ( cat tests.tar | xz -9e > tests.9e.tar.xz )
real 0m41.640s
user 0m41.009s
sys 0m0.602s
$ wc -c tests.9e.tar.xz
913432 tests.9e.tar.xz
After:
$ time ( cat tests.tar | xz -3 > tests.3.tar.xz )
real 0m2.422s
user 0m2.294s
sys 0m0.161s
$ wc -c tests.3.tar.xz
1556080 tests.3.tar.xz
Change-Id: Ia5c86183d4e19db3cbf95dd84cf86ee720ef4051
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
+ Add UDP_CPS, TCP_CPS, UDP_PPS and TCP_PPS suites.
+ Update existing cps traffic profiles.
+ Add missing traffic profiles.
+ UDP:
+ Single burst of 32 packets was confirmed as safe enough for TRex.
+ Maybe 64 could work, but not enough testing for that.
+ Multiple bursts have lead to reduced TRex performance,
as overlaping bursts (from different client instances)
tend to fill up the buffers.
+ TCP:
+ Data size set to 11111 bytes, completely arbitrarily.
+ Results look reasonable, so I have kept that.
- MSS not set at all
- No tested support for frame size other than 64B.
- Frame size does not even factor into TCP profiles.
+ So other frame sizes are skipped in autogen.
+ Update tags in related suites.
- HOSTS_{n} and SRC_USER_{n} should be unified.
- Questionable clarification on difference between IP4BASE and SCALE.
+ Add NAT state resetters to tests that need them.
+ Resetter is called (if set) before each measurement.
+ If ramp-up is detected, resetter is not set.
+ Rename "mult" argument to "multiplier".
+ Abstracted from packets to transactions.
+ Transaction corresponds to profile.
+ TRex multiplier argument sets target rate in transactions per second.
+ The familiar STL traffic:
+ Bidirectional is considered to be 2 packets per transaction.
+ Unidirectional is considered to be 1 packet per transaction.
+ The newer ASTF traffic:
+ 4 subtypes, each has different number of packets per transaction.
+ For max rate computation:
+ Packets in the more numerous direction are considered.
+ Rely on TRex reported traffic duration for ASTF:
+ Use the server side value.
- Client side value is higher by an overhead.
- TRex is not sending traffic during that time.
+ Remove delays from traffic profiles.
- Those delays would increase the reprted traffic time.
+ Support for scale lmited trials.
+ Only for ASTF profiles, each ASTF profile has limited scale.
+ Scale defined in suite variables.
+ For TRex to send all transactions provided duration value is ignored.
+ The appropriate value is computed in TrafficGenerator.
+ An ad-hoc time constant is added to match the TRex client side time overhead.
+ The profile driver receives the computed duration.
+ Measurement for PLRsearch add a sleep if the computed duration is smaller.
+ Alternative argument for search algos if scale is limited.
+ Both need higher timeout to accomodate big scales.
+ MLRsearch can afford fewer phases.
+ Added a parameter to optionally shorten the duration.
+ Use short duration for runtime stats trial and failure stats trial.
+ Use very large keepalive values in udp profiles to avoid ka packets.
+ No polling in ASTF profile driver.
- Polling could eliminate the time overhead value.
+ But polling proved to introduce some loss, affecting the results.
+ Handle duration stretching in ASTF by stopping traffic.
+ The stop has several steps so that:
+ The traffic is really stopped entirely.
+ Late packets do not count (maybe as errors).
+ Stats are preserved to read for results (and cleared afterwards).
+ Several quantities added to ReceiveRateMeasurement:
+ Original target duration is preserved (algos need that).
+ Input estimate (tps) for early search iterations.
+ Output estimate (maybe pps) for MRR output.
+ Strict result (unsent counts as loss) for NDR.
+ Use L2 counters (opackets, ipackets) where possible.
- TRex has trouble processing packets for the L7 ones at high loads.
+ Remove warmup from profile drivers and keywords.
+ Suites should call "Send ramp-up traffic" explicitly if needed.
+ Added parsing for few more counters.
+ Both to use in formulas or just for debug purposes.
- Only 64B cases in autogen, framesize support to be added later.
+ Latency streams during search can be enabled via PERF_USE_LATENCY env var.
+ MLRsearch improvments:
+ Rename argument names to min_rate and max_rate.
+ Use relative receive rate in initial phase.
+ PLRsearch improvements:
+ Careful computation when output (pps) does not match input (tps).
+ Use geometric distribution (instead of Poisson).
+ Helps agains math errors.
+ This should improve estimate stability.
- But in practice big losses still lead to significant jumps.
+ Traffic generator improvements:
+ send_traffic_on_tg now calls the full set_rate_provider_defaults.
+ _send_traffic_on_tg_internal for the logic without provider defaults.
+ As the internal function is re-used by measure() without affecting defaults.
+ Move _parse_traffic_results just before get_measurement_result.
+ As the latter uses fields set bu the former, it is now easier to read.
+ Multiple sources for approximate duration.
+ Tried from more precise to more available.
+ Includes logic for _pps tests (added in later change).
+ Move explicit type conversions to earlier occurences.
+ Profile driver output field uses semicolons to simplify parsing.
+ Performance Robot lib file split to several smaller ones.
+ performance_actions.robot:
+ Hosts Additional Statistics Action For * keywords.
+ performance_display.robot:
+ Hosts keyword for displaying and verifying results.
+ Change test message to use the correct unit (pps or cps).
+ performance_limits.robot renamed to performance_vars.robot
+ Added many keywords, mostly for accessing test variables.
+ Moved variables for Policer into a new keyword there.
+ Some keywords need sophisticated logic.
- Other are basically Get Variable Value.
+ But in future more logic can be added, without editing callers.
+ Documentation for the new keywords acts as a documentation for test variables.
+ performance_utils.robot has the rest.
+ Eliminated arguments if the value is in test variable.
+ Small improvements to documentation.
- Still not enough cleanup with respect to arguments and test variables.
+ Keywords are sorted alphabetically now in each one.
+ Suites:
+ Unified variables table:
+ No colons in comments.
+ ${n_hosts}, ${n_ports} and use them instead hardcoded numbers.
+ Add -cps to existing cps suite names.
+ Remove "trial data overwrite".
+ Compute max rate as in STL suites.
+ Each NAT suite has ip4base suite to compare results to.
- Those act as indirect TRex calibration.
- VPP does not lose packets in those.
+ Latency in ASTF suites is disabled hard.
- As we do not support latency in ASTF profiles yet.
+ Unidirectional tests governed by suite variable, not an argument.
+ Write long argument lists vertically.
+ Prefer to use argument names.
+ In Python, also the last argument is followed by comma.
+ It makes renaming and reordering easier.
+ Similarly applies to prints with long lists of values.
+ A TODO to update api crc file comments.
Change-Id: I84729355edbec051298a9de1162107f88ff5737d
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
Change-Id: I409b060f64ae7c6787448ae519fd76d8384e9ffb
Signed-off-by: Martin Balaz <martin.balaz@pantheon.tech>
|
|
- cover API changes in VPP: https://gerrit.fd.io/r/c/vpp/+/29463
- update vpp stable to version 21.01-rc0~283-g5f4f2081c
Change-Id: I079c10e4537448c3b078f22c3fe4ed266a5e2e2c
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
+ Not sure who was doing code.
+ Set interface up should have namespace version
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: I1103cddb9dd11e25a8ab0ce3f5126b2db9444749
|
|
There were occasional Router Advertisement packets received in a Traffic
Generator in VM testcases. Fix by adding config that disables sending
IPv6 RA messages on interfaces in VM VPP config before enabling them.
Also rename related keywords and change descriptions based on
https://docs.fd.io/vpp/21.01/db/d4c/clicmd_src_vnet_ip6-nd.html#clicmd_ip6_nd
Change-Id: I1c2a33337ac91f6039d287b4c2aac388e3a52383
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech>
|
|
- Specs
- Tests
- Library
- TODO: PAL
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: I8890940d56c7a9924d4a5f9365b6bb881610db14
|
|
Replace the hacky grep of /etc/resolv.conf with default docker gateway
IP.
Change-Id: Iec3a4658826f2ba871acb14d511e9c79a1273290
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech>
|
|
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: I3954157e91aafd370c3ac0843708903d19b84936
|
|
One ThunderX2 9975 server (.69) was replaced with two ThunderX2 9980
(.70, .71) servers. Move the .69 server under ansible perf section in
anticipation of repurposing it for that purpose. Update the ansible
scripts with .70 and .71 config and rename port names in device.sh lib
to reflect the NIC differences between .69 and .70 (and .71).
Change-Id: I88b75648735243e5559175d3192ffcc8fc70071c
Signed-off-by: Juraj Linkeš <juraj.linkes@pantheon.tech>
|
|
Change-Id: I3aa50ec1ef9b0445014daa31e767323060f4a03f
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
Change-Id: Ie24184ca4ac2d6c7abc32f0f103e10bc402ad93b
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
Change-Id: I1c638aef886bf37a9feb4a29e4949c7c8f19b717
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
- default value for main heap size set to 2G
- allow allocation of n x 2G in tests that need it
Change-Id: I02dafe6bc61649d2823486a3e983f66f5a213a88
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|
|
The previous 60 Mpps stil leads to ~30% duration stretching.
+ Add comment on why 36 Mpps was chosen as the new limit.
Change-Id: Ic11e8ece03939bdc8680cd7bc4122373583a2f17
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
- Follow up https://jira.fd.io/browse/VPP-1934
Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: Id0a26c5f67f229480332530a8531401d954f4422
|
|
- some tests need to reduce rate for ramp-up phase
- some tests need to extend trail duration in ramp-up phase
- removed 2n1l-10ge2p1x710-ethip4udp-nat44det-h1-p63-s63 suite
as nat out ports are randomly selected from available port range
so T-Rex stateless is not able to provide required out2in traffic
Change-Id: I1145496610d202f81d911e68aa819844d7600918
Signed-off-by: Jan Gelety <jgelety@cisco.com>
|