sonic-buildimage icon indicating copy to clipboard operation
sonic-buildimage copied to clipboard

[Chassis][multi-asic] Stopping swss services on SFMs should remove the related fabric reachability links but it does not

Open mannytaheri opened this issue 10 months ago • 2 comments

Description

When stopping swss services on SFMs, the related fabric reachability information do not get remove from the database.

Steps to reproduce the issue:

1.Ensure the related swss a syncd are up and running.

admin@ixre-cpm-chassis11:~$ docker ps -a
CONTAINER ID   IMAGE                                COMMAND                             CREATED      STATUS    PORTS     NAMES
1c6ac7cd1594   docker-syncd-brcm-dnx:latest     "/usr/local/bin/supeâ¦"   5 days ago   Up 27 hours            syncd15
2a0675724b5a   docker-orchagent:latest              "/usr/bin/docker-iniâ¦"   5 days ago   Up 27 hours             swss14
9f06a06e5907   docker-orchagent:latest               "/usr/bin/docker-iniâ¦"   5 days ago   Up 27 hours             swss15
7b6bde1b5460   docker-syncd-brcm-dnx:latest   "/usr/local/bin/supeâ¦"   5 days ago   Up 27 hours             syncd14

2.show fabric reachability information for acic14 and asic15

admin@ixre-cpm-chassis11:~$ show fabric reachability -n asic14 | grep up | wc 72 288 3960 admin@ixre-cpm-chassis11:~$ show fabric reachability -n asic15 | grep up | wc 72 288 3960

3.Stop swss services

admin@ixre-cpm-chassis11:~$ sudo systemctl stop swss@14 admin@ixre-cpm-chassis11:~$ sudo systemctl stop swss@15

  1. Check swss and syncd dockers. Ensure they have exited.
admin@ixre-cpm-chassis11:~$ docker ps -a
1c6ac7cd1594   docker-syncd-brcm-dnx:latest     "/usr/local/bin/supeâ¦"   5 days ago   Exited (0) 5 seconds ago     syncd15
2a0675724b5a   docker-orchagent:latest              "/usr/bin/docker-iniâ¦"   5 days ago   Exited (0) 26 seconds ago    swss14
9f06a06e5907   docker-orchagent:latest               "/usr/bin/docker-iniâ¦"   5 days ago   Exited (0) 11 seconds ago    swss15
7b6bde1b5460   docker-syncd-brcm-dnx:latest   "/usr/local/bin/supeâ¦"   5 days ago   Exited (0) 18 seconds ago    syncd14
  1. Show fabric reachibilty information on each related asics. These entries should have been removed but they are not.

admin@ixre-cpm-chassis11:~$ show fabric reachability -n asic15 | grep up | wc 72 288 3960 admin@ixre-cpm-chassis11:~$ show fabric reachability -n asic14 | grep up | wc 72 288 3960

Describe the results you received:

Describe the results you expected:

Output of show version:

(paste your output here)

Output of show techsupport:

(paste your output here or download and attach the file here )

Additional information you deem important (e.g. issue happens only occasionally):

mannytaheri avatar Apr 03 '24 14:04 mannytaheri

@deepak-singhal0408 @Aravind-Subbaroyan @mlok-nokia The information is in STATE_DB. such as key of FABRIC_PORT_TABLE|PORT100. Symptom is that of stop swss docker. The FABRIC_PORT_TABLE| corresponding ports STATUS are still up. Should some routine do update of tables in STATE_DB? Or can I add code in swss.sh in stop() case to cleanup the FABRIC_PORT_TABLE entries?

wumiaont avatar Apr 03 '24 20:04 wumiaont

  • @arlakshm @abdosi

deepak-singhal0408 avatar Apr 03 '24 21:04 deepak-singhal0408