Open wdoekes opened 2 weeks ago
Parts are fixed in master. But it's not backported to 202405, and it might bite again later.
Good fixes:
diff --git a/dockers/docker-database/critical_processes.j2 b/dockers/docker-database/critical_processes.j2 index 1f524132e..5bf4dbce4 100644 --- a/dockers/docker-database/critical_processes.j2 +++ b/dockers/docker-database/critical_processes.j2 @@ -1,5 +1,5 @@ -{% if INSTANCES %} -{% for redis_inst, redis_items in INSTANCES.items() %} +{% if INSTANCES -%} +{% for redis_inst, redis_items in INSTANCES.items() -%} program:{{ redis_inst }} -{%- endfor %} -{%- endif %} +{% endfor -%} +{% endif -%} diff --git a/dockers/docker-fpm-frr/frr/supervisord/critical_processes.j2 b/dockers/docker-fpm-frr/frr/supervisord/critical_processes.j2 index 69f4e8e69..4d65186b6 100644 --- a/dockers/docker-fpm-frr/frr/supervisord/critical_processes.j2 +++ b/dockers/docker-fpm-frr/frr/supervisord/critical_processes.j2 @@ -2,11 +2,11 @@ program:zebra program:staticd program:bgpd program:fpmsyncd -{% if DEVICE_METADATA.localhost.frr_mgmt_framework_config is defined and DEVICE_METADATA.localhost.frr_mgmt_framework_config == "true" %} +{% if DEVICE_METADATA.localhost.frr_mgmt_framework_config is defined and DEVICE_METADATA.localhost.frr_mgmt_framework_config == "true" -%} program:bfdd program:ospfd program:pimd program:frrcfgd -{%- else %} +{% else -%} program:bgpcfgd -{%- endif %} +{% endif -%} diff --git a/dockers/docker-orchagent/critical_processes.j2 b/dockers/docker-orchagent/critical_processes.j2 index b9bad7408..4b7f8accd 100644 --- a/dockers/docker-orchagent/critical_processes.j2 +++ b/dockers/docker-orchagent/critical_processes.j2 @@ -1,11 +1,11 @@ -{% set is_fabric_asic = 0 %} -{% if DEVICE_METADATA.localhost.switch_type %} -{% if DEVICE_METADATA.localhost.switch_type == "fabric" %} -{% set is_fabric_asic = 1 %} -{% endif %} -{% endif %} +{% set is_fabric_asic = 0 -%} +{% if DEVICE_METADATA.localhost.switch_type -%} +{% if DEVICE_METADATA.localhost.switch_type == "fabric" -%} +{% set is_fabric_asic = 1 -%} +{% endif -%} +{% endif -%} program:orchagent -{% if is_fabric_asic == 0 %} +{% if is_fabric_asic == 0 -%} program:portsyncd program:neighsyncd program:fdbsyncd @@ -19,4 +19,4 @@ program:nbrmgrd program:vxlanmgrd program:coppmgrd program:tunnelmgrd -{%- endif %} +{% endif %}
See #-20636.
Unsure if this is relevant. Looks new if this was not in the september-master-issue.
Looks like this:
[program:redis] command=/bin/bash -c "{ [[ -s /var/lib/redis/dump.rdb ]] || rm -f /var/lib/redis/dump.rdb; } && mkdir -p /var/lib/redis && exec /usr/bin/redis-server /etc/redis/redis.conf --bind 127.0.0.1 --port 6379 --unixsocket /var/run/redis/redis.sock --pidfile /var/run/redis/redis.pid --dir /var/lib/redis " priority=2 user=redis autostart=true autorestart=false stdout_logfile=syslog stderr_logfile=syslog
[program:redis_bmp] command=/bin/bash -c "{ [[ -s /var/lib/redis_bmp/dump.rdb ]] || rm -f /var/lib/redis_bmp/dump.rdb; } && mkdir -p /var/lib/redis_bmp && exec /usr/bin/redis-server /etc/redis/redis.conf --bind 127.0.0.1 --port 6400 --unixsocket /var/run/redis/redis_bmp.sock --pidfile /var/run/redis/redis_bmp.pid --dir /var/lib/redis_bmp " priority=2 user=redis autostart=true autorestart=false stdout_logfile=syslog stderr_logfile=syslog
^- this is likely since 06c469e6cc0a13beae45f2bdbea8c0d0ee82946c
Cause, probably this:
root@spine1:/# cat /etc/supervisor/critical_processes program:redis root@leaf1:/# cat /etc/supervisor/critical_processes program:redisprogram:redis_bmp
./dockers/docker-database/critical_processes.j2
{% if INSTANCES %} {% for redis_inst, redis_items in INSTANCES.items() %} program:{{ redis_inst }} {%- endfor %} {%- endif %}
Excess jinja line removal.
We'll do:
github.com/sonic-net/sonic-buildimage/pull 20897
And leave the rest.
The jinja stuff in sonig-cfggen is different and strip()s output mostly, so these fixes don't add any better output.
UPDATE
Parts are fixed in master. But it's not backported to 202405, and it might bite again later.
Good fixes:
ORIGINAL
See #-20636.
Unsure if this is relevant. Looks new if this was not in the september-master-issue.
Looks like this:
^- this is likely since 06c469e6cc0a13beae45f2bdbea8c0d0ee82946c
Cause, probably this:
./dockers/docker-database/critical_processes.j2
Excess jinja line removal.