Dualtor AA enablement for QOS SAI#17765
Conversation
|
/azp run |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
PR test failed. Re-try. |
| - https://github.com/sonic-net/sonic-mgmt/issues/12906 | ||
| - "topo_type in ['m0', 'mx', 'm1', 'm2', 'm3']" | ||
| - "topo_name not in ['t0', 't0-64', 't0-116', 't0-118', 't0-35', 't0-56', 't0-standalone-32', 't0-standalone-64', 't0-standalone-128', 't0-standalone-256', 'dualtor-56', 'dualtor-120', 'dualtor', 't0-80', 't0-backend', 't1-lag', 't1-28-lag', 't1-64-lag', 't1-56-lag', 't1-backend', 't2', 't2_2lc_36p-masic', 't2_2lc_min_ports-masic'] and asic_type not in ['mellanox']" | ||
| - "topo_name not in ['t0', 't0-64', 't0-116', 't0-118', 't0-35', 't0-56', 't0-standalone-32', 't0-standalone-64', 't0-standalone-128', 't0-standalone-256', 'dualtor-56', 'dualtor-120', 'dualtor', 'dualtor-aa', 't0-80', 't0-backend', 't1-lag', 't1-28-lag', 't1-64-lag', 't1-56-lag', 't1-backend', 't2', 't2_2lc_36p-masic', 't2_2lc_min_ports-masic'] and asic_type not in ['mellanox']" |
There was a problem hiding this comment.
missing dualtor-aa-56, dualtor-aa-120, dualtor-aa-64, dualtor-aa-64-breakout?
There was a problem hiding this comment.
Some discussion led us to coordinate on dualtor-aa, and that is the topo I've tested on. If you can confirm you'd like the rest of the dualtor-aa series I can file a PR.
This duplicated list should also be fixed, my proposed fix is #17707 . If there's an alternative for QOS SAI it should be filed.
There was a problem hiding this comment.
The duplication is now resulting in duplicated merge conflict issues. This needs to be resolved.
| is_dualtor_active_standby = is_dualtor and active_standby_ports | ||
| """ Stop mux container for dual ToR Active-Standby """ | ||
| if is_dualtor: | ||
| if is_dualtor_active_standby: |
There was a problem hiding this comment.
if the dualtor is active-active, the practice is to set the active-active tors pair into active-standby so all upstream packets will only be forwarded to the active side; otherwise, the upstream packets will be ECMPed.
Please refer to #16549.
There was a problem hiding this comment.
This does achieve putting the upper tor into standby state, though it'd be preferrable to use something much more simple as in 16549 and eliminate a lot of the current AS code.
The fixtures in 16549 default to function scope, and QOS SAI needs a class scope. So the change is a bit non-trivial, requiring some more thorough testing.
|
@kevinskwang should we cherry-pick to 202411? |
* Apply diff for dualtor AA QOS SAI with list change duplicated for all qos sai topo lists. * Add back a newline.
Code sync sonic-net/sonic-mgmt:202411 => 202412 ``` * 86163d9 (HEAD -> code-sync-202412, origin/code-sync-202412) r12f 250620:0224 - Merge remote-tracking branch 'base/202411' into code-sync-202412 |\ | * 6c7825f (base/202411) Justin Wong 250611:2109 - Skip test_bgp_multipath_relax.py on t1-isolated (sonic-net#18558) | * 829aa40 Chun'ang Li 250619:1618 - [CI][202411]update pr test template reference (sonic-net#19059) | * 6f2a1bd prabhataravind 250617:1800 - Fix port mapping for new 4280 hwskus (sonic-net#18802) (sonic-net#19030) | * 66c7667 ganglv 250617:1623 - [202411] Update timeout for gnmi subscribe API (sonic-net#19045) | * 8f0a97b Justin Wong 250611:2146 - Extend bgp command check to allow multipath routes and "best" routes (sonic-net#18910) | * 12b7793 StormLiangMS 250611:1110 - [XFAIL: test_crm_available.py] set xfail of test_crm_available on dualtor platform (sonic-net#18874) | * ab5fd70 rbpittman 250615:2148 - Shutdown BGP instead of stopping some processes. (sonic-net#18267) | * 3f35fa2 Harish Kalyanaraman 250321:0332 - Increase GCUTIMEOUT value for nokia-armhf platforms (sonic-net#17420) | * 48bf447 xwjiang-ms 250613:1238 - [202411] Support Ubuntu 24 server in KVM (sonic-net#18929) | * 194b907 rbpittman 250612:2057 - Dualtor AA enablement for QOS SAI (sonic-net#17765) (sonic-net#17911) ```
* Apply diff for dualtor AA QOS SAI with list change duplicated for all qos sai topo lists. * Add back a newline. Signed-off-by: opcoder0 <[email protected]>
* Apply diff for dualtor AA QOS SAI with list change duplicated for all qos sai topo lists. * Add back a newline. Signed-off-by: Guy Shemesh <[email protected]>
* Apply diff for dualtor AA QOS SAI with list change duplicated for all qos sai topo lists. * Add back a newline. Signed-off-by: Guy Shemesh <[email protected]>
Description of PR
Summary:
Type of change
Back port request
Approach
What is the motivation for this PR?
How did you do it?
How did you verify/test it?
Ran on dualtor AA cisco-8000, verified tests execute without major errors. Some tests fail due to some normal instability and known issues.
Any platform specific information?
Supported testbed topology if it's a new test case?
Documentation