[RouteOrch] Record ROUTE_TABLE entry programming status to APPL_STATE_DB by stepanblyschak · Pull Request #2512 · sonic-net/sonic-swss

stepanblyschak · 2022-11-11T13:39:42Z

Signed-off-by: Stepan Blyschak stepanb@nvidia.com

DEPENDS ON #2511

What I did

I added ROUTE_TABLE entry programming status recording to APPL_STATE_DB. I made a ResponsePublisher work in buffered mode to be more optimized for route bulk processing.

I always publish a route entry with return code SAI_STATUS_SUCCESS because orchagent immidiatelly exits on any error.

Why I did it

I did it to support route supression in BGP. Orchagent has to sends route programming feedback back to fpmsyncd which communicates with zebra.

How I verified it

I wrote UT, also on the switch, you can run:

admin@sonic:~$ redis-cli -n 14 keys '*ROUTE*'

Details if related

Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

prsunny · 2023-01-10T20:00:14Z

orchagent/routeorch.cpp

                    if (fvField(i) == "seg_src")
                        srv6_source = fvValue(i);
+
+                    if (fvField(i) == "protocol")


Why do we have to introduce this field?

It is required to construct RTM_NEWROUTE with correct proto number.
https://github.com/stepanblyschak/SONiC/blob/bgp-suppress-fib-pending/doc/BGP/BGP-supress-fib-pending.md#74-fpmsyncd

prsunny · 2023-01-10T20:04:50Z

orchagent/routeorch.cpp

+    /* If the operation type is "DEL" then ctx.protocol is empty and fvs is left empty.
+     * An empty fvs makes ResponsePublisher::publish() remove the state entry from APPL_STATE_DB
+     */
+    if (!ctx.protocol.empty())


Why introduce this field for identifying operation? I'm not sure if i understand the logic here. Can you please explaing this?

It is not the motivation for introducing this field. The context does not carry operation type, however, if the protocol is left empty that means it is a delete operation, since the delete update does not carry FVS (including protocol)

@prsunny Replaced with a flag saved in the context object, please check.

Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

…k/sonic-swss into routeorch-response-channel

prsunny

lgtm,

This is needed in case DEL/SET consolidation by ConsumerStateTable Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

prsunny · 2023-02-09T01:23:24Z

/azp run

azure-pipelines · 2023-02-09T01:23:34Z

Azure Pipelines successfully started running 1 pipeline(s).

prsunny · 2023-02-09T01:27:45Z

tests/mock_tests/Makefile.am

 tests_intfmgrd_CPPFLAGS = $(DBGFLAGS) $(AM_CFLAGS) $(CFLAGS_COMMON) $(CFLAGS_GTEST) $(CFLAGS_SAI) $(tests_intfmgrd_INCLUDES)
 tests_intfmgrd_LDADD = $(LDADD_GTEST) $(LDADD_SAI) -lnl-genl-3 -lhiredis -lhiredis \
-        -lswsscommon -lswsscommon -lgtest -lgtest_main -lzmq -lnl-3 -lnl-route-3 -lpthread
+        -lswsscommon -lswsscommon -lgtest -lgtest_main -lzmq -lnl-3 -lnl-route-3 -lpthread -lgmock -lgmock_main


why do we need this change in intfmgrd?

@prsunny tests/mock_tests/fake_response_publisher.cpp tests_intfmgrd depends on is using google mock

stepanblyschak · 2023-02-10T09:09:43Z

/azpw run

mssonicbld · 2023-02-10T09:09:44Z

/AzurePipelines run

azure-pipelines · 2023-02-10T09:09:55Z

Azure Pipelines successfully started running 1 pipeline(s).

Update sonic-swss submodule pointer to include the following: * f66abed Support for tc-dot1p and tc-dscp qosmap ([sonic-net#2559](sonic-net/sonic-swss#2559)) * 35385ad [RouteOrch] Record ROUTE_TABLE entry programming status to APPL_STATE_DB ([sonic-net#2512](sonic-net/sonic-swss#2512)) * 0704f78 [Workaround] EvpnRemoteVnip2pOrch warmboot check failure ([sonic-net#2626](sonic-net/sonic-swss#2626)) * 4df5cab [ResponsePublisher] add pipeline support ([sonic-net#2511](sonic-net/sonic-swss#2511)) Signed-off-by: AntonHryshchuk <antonh@nvidia.com>

Update sonic-swss submodule pointer to include the following: * baa302e Do not allow to add port to .1Q bridge while router port deletion is not completed ([sonic-net#2669](sonic-net/sonic-swss#2669)) * f66abed Support for tc-dot1p and tc-dscp qosmap ([sonic-net#2559](sonic-net/sonic-swss#2559)) * 35385ad [RouteOrch] Record ROUTE_TABLE entry programming status to APPL_STATE_DB ([sonic-net#2512](sonic-net/sonic-swss#2512)) * 0704f78 [Workaround] EvpnRemoteVnip2pOrch warmboot check failure ([sonic-net#2626](sonic-net/sonic-swss#2626)) * 4df5cab [ResponsePublisher] add pipeline support ([sonic-net#2511](sonic-net/sonic-swss#2511)) Signed-off-by: dprital <drorp@nvidia.com>

Update sonic-swss submodule pointer to include the following: * baa302e Do not allow to add port to .1Q bridge while router port deletion is not completed ([#2669](sonic-net/sonic-swss#2669)) * f66abed Support for tc-dot1p and tc-dscp qosmap ([#2559](sonic-net/sonic-swss#2559)) * 35385ad [RouteOrch] Record ROUTE_TABLE entry programming status to APPL_STATE_DB ([#2512](sonic-net/sonic-swss#2512)) * 0704f78 [Workaround] EvpnRemoteVnip2pOrch warmboot check failure ([#2626](sonic-net/sonic-swss#2626)) * 4df5cab [ResponsePublisher] add pipeline support ([#2511](sonic-net/sonic-swss#2511))

DEPENDS: #2512 What I did I implemented support to enable pending routes suppression feature. When this feature is enabled, fpmsyncd will wait for reply from orchagent before sending offload status message to zebra. Why I did it This is done to not announce routes which aren't yet offloaded in HW. How I verified it UT and manual tests.

…PL_STATE_DB (sonic-net#2512)" This reverts commit 35385ad.

* Revert "[RouteOrch] Record ROUTE_TABLE entry programming status to APPL_STATE_DB (#2512)" This reverts commit 35385ad. * Revert "[fpmsyncd] Implement pending route suppression feature (#2551)" This reverts commit 840fe1d. * Get back gMockResponsePublisher as it is used by other tests Signed-off-by: Stepan Blyschak <stepanb@nvidia.com> --------- Signed-off-by: Stepan Blyschak <stepanb@nvidia.com> Co-authored-by: Ying Xie <yxieca@users.noreply.github.com>

…PL_STATE_DB (sonic-net#2512)" This reverts commit 35385ad.

…_DB (sonic-net#2512) * [RouteOrch] Record ROUTE_TABLE entry programming status to APPL_STATE_DB * Support route supression in BGP. Orchagent has to sends route programming feedback back to fpmsyncd which communicates with zebra.

DEPENDS: sonic-net#2512 What I did I implemented support to enable pending routes suppression feature. When this feature is enabled, fpmsyncd will wait for reply from orchagent before sending offload status message to zebra. Why I did it This is done to not announce routes which aren't yet offloaded in HW. How I verified it UT and manual tests.

stepanblyschak added 2 commits November 8, 2022 15:36

[RouteOrch] Record ROUTE_TABLE entry programming status to APPL_STATE_DB

99dd9e4

Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

fix comment

44d1c21

Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

stepanblyschak requested a review from prsunny as a code owner November 11, 2022 13:39

This was referenced Nov 11, 2022

[config/show] Add command to control pending FIB suppression sonic-net/sonic-utilities#2495

Merged

FIB Suppress Announcements of Routes Not Installed in HW sonic-net/SONiC#1103

Merged

stepanblyschak added 2 commits November 17, 2022 14:44

set replace false

47a65eb

Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

set test expect replace=false

e9feb5e

Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

stepanblyschak force-pushed the routeorch-response-channel branch from a4b7a55 to e9feb5e Compare November 17, 2022 19:18

Merge branch 'master' into routeorch-response-channel

74fd843

liat-grozovik requested a review from marian-pritsak November 28, 2022 13:53

stepanblyschak mentioned this pull request Nov 29, 2022

[fpmsyncd] Implement pending route suppression feature #2551

Merged

prsunny reviewed Jan 10, 2023

View reviewed changes

stepanblyschak added 2 commits January 11, 2023 15:53

use flag to determite operation type

e50d5cd

Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

Merge branch 'routeorch-response-channel' of github.com:stepanblyscha…

2b80c52

…k/sonic-swss into routeorch-response-channel

prsunny previously approved these changes Jan 12, 2023

View reviewed changes

prsunny requested a review from qiluo-msft January 12, 2023 21:58

publish route install status in case a duplicate prefix is received

0fa58de

This is needed in case DEL/SET consolidation by ConsumerStateTable Signed-off-by: Stepan Blyschak <stepanb@nvidia.com>

stepanblyschak dismissed prsunny’s stale review via 0fa58de January 19, 2023 18:00

prsunny reviewed Feb 9, 2023

View reviewed changes

prsunny approved these changes Feb 14, 2023

View reviewed changes

prsunny merged commit 35385ad into sonic-net:master Feb 14, 2023

dprital mentioned this pull request Feb 23, 2023

[submodule] Advance sonic-swss pointer sonic-net/sonic-buildimage#13953

Merged

8 tasks

stepanblyschak added a commit to stepanblyschak/sonic-swss that referenced this pull request Dec 19, 2023

Revert "[RouteOrch] Record ROUTE_TABLE entry programming status to AP…

5def21c

…PL_STATE_DB (sonic-net#2512)" This reverts commit 35385ad.

stepanblyschak added a commit to stepanblyschak/sonic-swss that referenced this pull request Dec 20, 2023

Revert "[RouteOrch] Record ROUTE_TABLE entry programming status to AP…

95f24e2

…PL_STATE_DB (sonic-net#2512)" This reverts commit 35385ad.

stepanblyschak added a commit to stepanblyschak/sonic-swss that referenced this pull request May 17, 2024

Revert "[RouteOrch] Record ROUTE_TABLE entry programming status to AP…

a4f2bd8

…PL_STATE_DB (sonic-net#2512)" This reverts commit 35385ad.

Conversation

stepanblyschak commented Nov 11, 2022

Uh oh!

prsunny Jan 10, 2023

Choose a reason for hiding this comment

Uh oh!

stepanblyschak Jan 11, 2023

Choose a reason for hiding this comment

Uh oh!

prsunny Jan 10, 2023

Choose a reason for hiding this comment

Uh oh!

stepanblyschak Jan 11, 2023

Choose a reason for hiding this comment

Uh oh!

stepanblyschak Jan 11, 2023

Choose a reason for hiding this comment

Uh oh!

prsunny left a comment

Choose a reason for hiding this comment

Uh oh!

prsunny commented Feb 9, 2023

Uh oh!

azure-pipelines bot commented Feb 9, 2023

Uh oh!

prsunny Feb 9, 2023

Choose a reason for hiding this comment

Uh oh!

stepanblyschak Feb 9, 2023

Choose a reason for hiding this comment

Uh oh!

stepanblyschak commented Feb 10, 2023

Uh oh!

mssonicbld commented Feb 10, 2023

Uh oh!

azure-pipelines bot commented Feb 10, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants