Next Meeting: TBD (will be announced on the ci@dpdk.org mailing list)
To join the Community CI meetings, please send an email to Trishan de Lanerolle <tdelanerolle@linuxfoundation.org>.
Lincoln Lavoie, UNH-IOL
Bradon Lo, UNH-IOL
David Liu, UNH-IOL
Aaron Conole
Cathal O’Hare
Gal Cohen
John McNamara
Thomasz Zawadzki
Trishan De Lanerolle
Ali Alnubani
Thomas Monjalon
Meeting Schedule & Logistics & Minutes
CI Status
Bugzilla Ticket Review
Any other business
Schedule & Duration
Keep to every other week schedule
Lincoln will confirm the time slot with James
Minutes
Need to finalize a location to keep minutes that is accessible to all participant
Minutes will be sent to the ci@dpdk.com mailing list, can put the
Add information for who to contact to join the meetings and get the calendar invite, etc.
UNH hosted community lab
Patched false failures based on processor speed, instead of past run results
NXP has been contacted to keep an eye on false failures
Mellanox run failures being investigated, no common root cause determined yet
May be caused by TRex failures / linking with DPDK fork
Multiple emails for one context in patchworks has been fixed, and should prevent “last update” email from causing false pass or fail.
Still working with Broadcom, should be completed and online within the next two weeks.
Travis CI
Multi-arch support is a little buggy
Trishan will contact Travis CI to find out what would be gained from a paid account, and if they would eliminate the disk quota limitation
Other CI Tools
Cirrus CI - could provide support for Linux, BSD, and Window
Aaron will take a look at this, and report back what would required to begin using this for DPDK
Github Actions -
SUSE Open Build Service - may be allowed to use for unit testing
ARM is looking at possible integrating this into DPDK patchwork
Other traffic generator tooling:
TRex forks DPDK, which should not be promoted. Can we help encourage them to “not fork.”
Should other open source tools be tried, i.e. Warp-L7, these need to be supported into DTS to be used in the current performance lab.
Redhat may be able to help investigate this, based on the past developers being hired into Redhat
Bug 425 - Eth Tests - Link Status Checks
Currently in code audit, to verify it’s not a duplicate
Bug 426 - Eth Tests - MTU Checks
Currently in code audit, to verify it’s not a duplicate
Bug 427 - Eth Tests - Promiscuous Support
Currently debugging the implementation
Bug 451 - Split dpdk-ci and dpdklab-ci scripts
Being started this week, since other bugs have been cleared out on the infrastructure.
Bug 459 - Performance Tuning and Results Stability Issues on NXP test environment
Being tracked, but we have not seen failures since the last meetings. Nothing has been changed on the systems since the past meeting.
Bug 428 - Eth Tests - Hardware Checksum Checks
Need to discuss this on the lists, including the API maintainers
What is to be tested?
Bug 210 - Patches with dependency fails to apply
Brandon will give an estimate on difficulty to implement (i.e. time), may change the priority based on this outcome.
None