sonic-buildimage/files/build_templates/docker_image_ctl.j2

621 lines
25 KiB
Plaintext
Raw Normal View History

#!/bin/bash
# single instance containers are still supported (even though it might not look like it)
# if no instance number is passed to this script, $DEV will simply be unset, resulting in docker
# commands being sent to the base container name. E.g. `docker start database$DEV` simply starts
# the container `database` if no instance number is passed since `$DEV` is not defined
{%- if docker_container_name == "database" %}
link_namespace() {
# Makes namespace of a docker container available in
# /var/run/netns so it can be managed with iproute2
mkdir -p /var/run/netns
PID="$(docker inspect -f {{"'{{.State.Pid}}'"}} "${DOCKERNAME}")"
PIDS=`ip netns pids "$NET_NS" 2>/dev/null`
if [ "$?" -eq "0" ]; then # namespace exists
if `echo $PIDS | grep --quiet -w $PID`; then # namespace is correctly linked
return 0
else # if it's incorrectly linked remove it
ip netns delete $NET_NS
fi
fi
ln -s /proc/$PID/ns/net /var/run/netns/$NET_NS
}
{%- endif %}
function updateSyslogConf()
{
# On multiNPU platforms, change the syslog target ip to docker0 ip to allow logs from containers
# running on the namespace to reach the rsyslog service running on the host
# Also update the container name
if [[ ($NUM_ASIC -gt 1) ]]; then
TARGET_IP=$(docker network inspect bridge --format={{ "'{{(index .IPAM.Config 0).Gateway}}'" }})
CONTAINER_NAME="$DOCKERNAME"
TMP_FILE="/tmp/rsyslog.$CONTAINER_NAME.conf"
{%- if docker_container_name == "database" %}
python -c "import jinja2, os; paths=['/usr/share/sonic/templates']; loader = jinja2.FileSystemLoader(paths); env = jinja2.Environment(loader=loader, trim_blocks=True); template_file='/usr/share/sonic/templates/rsyslog-container.conf.j2'; template = env.get_template(os.path.basename(template_file)); data=template.render({\"target_ip\":\"$TARGET_IP\",\"container_name\":\"$CONTAINER_NAME\"}); print(data)" > $TMP_FILE
{%- else %}
sonic-cfggen -t /usr/share/sonic/templates/rsyslog-container.conf.j2 -a "{\"target_ip\": \"$TARGET_IP\", \"container_name\": \"$CONTAINER_NAME\" }" > $TMP_FILE
{%- endif %}
docker cp $TMP_FILE ${DOCKERNAME}:/etc/rsyslog.conf
rm -rf $TMP_FILE
fi
}
function ebtables_config()
{
if [ "$DEV" ]; then
# Install ebtables filter in namespaces on multi-asic.
ip netns exec $NET_NS ebtables-restore < /etc/ebtables.filter.cfg
else
if [[ ! ($NUM_ASIC -gt 1) ]]; then
# Install ebtables filter in host for single asic.
ebtables-restore < /etc/ebtables.filter.cfg
fi
fi
}
function getMountPoint()
{
echo $1 | python -c "import sys, json, os; mnts = [x for x in json.load(sys.stdin)[0]['Mounts'] if x['Destination'] == '/usr/share/sonic/hwsku']; print('' if len(mnts) == 0 else os.path.abspath(mnts[0]['Source']))" 2>/dev/null
}
function getBootType()
{
# same code snippet in files/scripts/syncd.sh
case "$(cat /proc/cmdline)" in
*SONIC_BOOT_TYPE=warm*)
TYPE='warm'
;;
*SONIC_BOOT_TYPE=fastfast*)
TYPE='fastfast'
;;
*SONIC_BOOT_TYPE=fast*|*fast-reboot*)
TYPE='fast'
;;
*)
TYPE='cold'
esac
echo "${TYPE}"
}
function preStartAction()
{
{%- if docker_container_name == "database" %}
WARM_DIR=/host/warmboot
if [ "$DATABASE_TYPE" != "chassisdb" ]; then
if [[ ("$BOOT_TYPE" == "warm" || "$BOOT_TYPE" == "fastfast") && -f $WARM_DIR/dump.rdb ]]; then
# Load redis content from /host/warmboot/dump.rdb
docker cp $WARM_DIR/dump.rdb database$DEV:/var/lib/redis/dump.rdb
else
# Create an emtpy file and overwrite any RDB if already there
echo -n > /tmp/dump.rdb
docker cp /tmp/dump.rdb database$DEV:/var/lib/redis/
fi
fi
{%- elif docker_container_name == "pde" %}
if [[ $(/bin/systemctl status swss | grep -c running) -gt 0 ]]; then
echo "Stopping SWSS before starting PDE"
systemctl stop swss
fi
if [[ $(/bin/systemctl status watchdog-control.service | grep -c running) -gt 0 ]]; then
echo "Stopping watchdog-control.service before starting PDE"
systemctl stop watchdog-control.service
fi
{%- elif docker_container_name == "snmp" %}
$SONIC_DB_CLI STATE_DB HSET 'DEVICE_METADATA|localhost' chassis_serial_number $(decode-syseeprom -s)
{%- else %}
: # nothing
{%- endif %}
updateSyslogConf
}
{%- if docker_container_name == "database" %}
function setPlatformLagIdBoundaries()
{
[chassis] Fix issues regarding database service failure handling and mid-plane connectivity for namespace. (#10500) What/Why I did: Issue1: By setting up of ipvlan interface in interface-config.sh we are not tolerant to failures. Reason being interface-config.service is one-shot and do not have restart capability. Scenario: For example if let's say database service goes in fail state then interface-services also gets failed because of dependency check but later database service gets restart but interface service will remain in stuck state and the ipvlan interface nevers get created. Solution: Moved all the logic in database service from interface-config service which looks more align logically also since the namespace is created here and all the network setting (sysctl) are happening here.With this if database starts we recreate the interface. Issue 2: Use of IPVLAN vs MACVLAN Currently we are using ipvlan mode. However above failure scenario is not handle correctly by ipvlan mode. Once the ipvlan interface is created and ip address assign to it and if we restart interface-config or database (new PR) service Linux Kernel gives error "Error: Address already assigned to an ipvlan device." based on this:https://github.com/torvalds/linux/blob/master/drivers/net/ipvlan/ipvlan_main.c#L978Reason being if we do not do cleanup of ip address assignment (need to be unique for IPVLAN) it remains in Kernel Database and never goes to free pool even though namespace is deleted. Solution: Considering this hard dependency of unique ip macvlan mode is better for us and since everything is managed by Linux Kernel and no dependency for on user configured IP address. Issue3: Namespace database Service do not check reachability to Supervisor Redis Chassis Server. Currently there is no explicit check as we never do Redis PING from namespace to Supervisor Redis Chassis Server. With this check it's possible we will start database and all other docker even though there is no connectivity and will hit the error/failure late in cycle Solution: Added explicit PING from namespace that will check this reachability. Issue 4:flushdb give exception when trying to accces Chassis Server DB over Unix Sokcet. Solution: Handle gracefully via try..except and log the message.
2022-05-24 18:54:12 -05:00
docker exec -i ${DOCKERNAME} $SONIC_DB_CLI CHASSIS_APP_DB SET "SYSTEM_LAG_ID_START" "$lag_id_start"
docker exec -i ${DOCKERNAME} $SONIC_DB_CLI CHASSIS_APP_DB SET "SYSTEM_LAG_ID_END" "$lag_id_end"
}
function waitForAllInstanceDatabaseConfigJsonFilesReady()
{
if [ ! -z "$DEV" ]; then
cnt=0
SONIC_DB_GLOBAL_JSON="/var/run/redis/sonic-db/database_global.json"
if [ -f "$SONIC_DB_GLOBAL_JSON" ]; then
# Create a separate python script to get a list of location of all instance database_config.json file
redis_database_cfg_list=`/usr/bin/python -c "import sys; import os; import json; f=open(sys.argv[1]); \
global_db_dir = os.path.dirname(sys.argv[1]); data=json.load(f); \
print(\" \".join([os.path.normpath(global_db_dir+'/'+elem['include']) \
for elem in data['INCLUDES'] if 'namespace' in elem])); f.close()" $SONIC_DB_GLOBAL_JSON`
for file in $redis_database_cfg_list
do
while [ ! -f $file ]
do
sleep 1
cnt=$(( $cnt + 1))
if [ $cnt -ge 60 ]; then
echo "Error: $file not found"
break
fi
done
done
fi
fi
}
# delay a second to allow the file to be fully accessible
sleep 1
{%- endif %}
function postStartAction()
{
{%- if docker_container_name == "database" %}
[chassis] Fix issues regarding database service failure handling and mid-plane connectivity for namespace. (#10500) What/Why I did: Issue1: By setting up of ipvlan interface in interface-config.sh we are not tolerant to failures. Reason being interface-config.service is one-shot and do not have restart capability. Scenario: For example if let's say database service goes in fail state then interface-services also gets failed because of dependency check but later database service gets restart but interface service will remain in stuck state and the ipvlan interface nevers get created. Solution: Moved all the logic in database service from interface-config service which looks more align logically also since the namespace is created here and all the network setting (sysctl) are happening here.With this if database starts we recreate the interface. Issue 2: Use of IPVLAN vs MACVLAN Currently we are using ipvlan mode. However above failure scenario is not handle correctly by ipvlan mode. Once the ipvlan interface is created and ip address assign to it and if we restart interface-config or database (new PR) service Linux Kernel gives error "Error: Address already assigned to an ipvlan device." based on this:https://github.com/torvalds/linux/blob/master/drivers/net/ipvlan/ipvlan_main.c#L978Reason being if we do not do cleanup of ip address assignment (need to be unique for IPVLAN) it remains in Kernel Database and never goes to free pool even though namespace is deleted. Solution: Considering this hard dependency of unique ip macvlan mode is better for us and since everything is managed by Linux Kernel and no dependency for on user configured IP address. Issue3: Namespace database Service do not check reachability to Supervisor Redis Chassis Server. Currently there is no explicit check as we never do Redis PING from namespace to Supervisor Redis Chassis Server. With this check it's possible we will start database and all other docker even though there is no connectivity and will hit the error/failure late in cycle Solution: Added explicit PING from namespace that will check this reachability. Issue 4:flushdb give exception when trying to accces Chassis Server DB over Unix Sokcet. Solution: Handle gracefully via try..except and log the message.
2022-05-24 18:54:12 -05:00
CHASSISDB_CONF="/usr/share/sonic/device/$PLATFORM/chassisdb.conf"
[ -f $CHASSISDB_CONF ] && source $CHASSISDB_CONF
if [ "$DEV" ]; then
# Enable the forwarding on eth0 interface in namespace.
SYSCTL_NET_CONFIG="/etc/sysctl.d/sysctl-net.conf"
docker exec -i database$DEV sed -i -e "s/^net.ipv4.conf.eth0.forwarding=0/net.ipv4.conf.eth0.forwarding=1/;
s/^net.ipv6.conf.eth0.forwarding=0/net.ipv6.conf.eth0.forwarding=1/" $SYSCTL_NET_CONFIG
docker exec -i database$DEV sysctl --system -e
link_namespace $DEV
[chassis] Fix issues regarding database service failure handling and mid-plane connectivity for namespace. (#10500) What/Why I did: Issue1: By setting up of ipvlan interface in interface-config.sh we are not tolerant to failures. Reason being interface-config.service is one-shot and do not have restart capability. Scenario: For example if let's say database service goes in fail state then interface-services also gets failed because of dependency check but later database service gets restart but interface service will remain in stuck state and the ipvlan interface nevers get created. Solution: Moved all the logic in database service from interface-config service which looks more align logically also since the namespace is created here and all the network setting (sysctl) are happening here.With this if database starts we recreate the interface. Issue 2: Use of IPVLAN vs MACVLAN Currently we are using ipvlan mode. However above failure scenario is not handle correctly by ipvlan mode. Once the ipvlan interface is created and ip address assign to it and if we restart interface-config or database (new PR) service Linux Kernel gives error "Error: Address already assigned to an ipvlan device." based on this:https://github.com/torvalds/linux/blob/master/drivers/net/ipvlan/ipvlan_main.c#L978Reason being if we do not do cleanup of ip address assignment (need to be unique for IPVLAN) it remains in Kernel Database and never goes to free pool even though namespace is deleted. Solution: Considering this hard dependency of unique ip macvlan mode is better for us and since everything is managed by Linux Kernel and no dependency for on user configured IP address. Issue3: Namespace database Service do not check reachability to Supervisor Redis Chassis Server. Currently there is no explicit check as we never do Redis PING from namespace to Supervisor Redis Chassis Server. With this check it's possible we will start database and all other docker even though there is no connectivity and will hit the error/failure late in cycle Solution: Added explicit PING from namespace that will check this reachability. Issue 4:flushdb give exception when trying to accces Chassis Server DB over Unix Sokcet. Solution: Handle gracefully via try..except and log the message.
2022-05-24 18:54:12 -05:00
if [[ -n "$midplane_subnet" ]]; then
# Use /16 for loopback interface
ip netns exec "$NET_NS" ip addr add 127.0.0.1/16 dev lo
ip netns exec "$NET_NS" ip addr del 127.0.0.1/8 dev lo
# Create eth1 in database instance
ip link add name ns-eth1"$NET_NS" link eth1-midplane type macvlan mode bridge
ip link set dev ns-eth1"$NET_NS" netns "$NET_NS"
ip netns exec "$NET_NS" ip link set ns-eth1"$NET_NS" name eth1
# Configure IP address and enable eth1
lc_slot_id=$(python3 -c 'import sonic_platform.platform; platform_chassis = sonic_platform.platform.Platform().get_chassis(); print(platform_chassis.get_my_slot())' 2>/dev/null)
lc_ip_address=`echo $midplane_subnet | awk -F. '{print $1 "." $2}'`.$lc_slot_id.$(($DEV + 10))
lc_subnet_mask=${midplane_subnet#*/}
ip netns exec "$NET_NS" ip addr add $lc_ip_address/$lc_subnet_mask dev eth1
ip netns exec "$NET_NS" ip link set dev eth1 up
# Allow localnet routing on the new interfaces if midplane is using a
# subnet in the 127/8 range.
if [[ "${midplane_subnet#127}" != "$midplane_subnet" ]]; then
ip netns exec "$NET_NS" bash -c "echo 1 > /proc/sys/net/ipv4/conf/eth1/route_localnet"
fi
fi
fi
# Setup ebtables configuration
{%- if sonic_asic_platform != "vs" %}
ebtables_config
{%- endif %}
# chassisdb starts before database starts, bypass the PING check since other
# databases are not availbale until database container is ready.
# also chassisdb doesn't support warm/fast reboot, its dump.rdb is deleted
# at service startup time, nothing need to be done here.
if [ "$DATABASE_TYPE" != "chassisdb" ]; then
# Wait until supervisord and redis starts. This change is needed
# because now database_config.json is jinja2 templated based
# and by the time file gets generated if we do redis ping
# then we catch python exception of file not valid
# that comes to syslog which is unwanted so wait till database
# config is ready and then ping
[chassis] Fix issues regarding database service failure handling and mid-plane connectivity for namespace. (#10500) What/Why I did: Issue1: By setting up of ipvlan interface in interface-config.sh we are not tolerant to failures. Reason being interface-config.service is one-shot and do not have restart capability. Scenario: For example if let's say database service goes in fail state then interface-services also gets failed because of dependency check but later database service gets restart but interface service will remain in stuck state and the ipvlan interface nevers get created. Solution: Moved all the logic in database service from interface-config service which looks more align logically also since the namespace is created here and all the network setting (sysctl) are happening here.With this if database starts we recreate the interface. Issue 2: Use of IPVLAN vs MACVLAN Currently we are using ipvlan mode. However above failure scenario is not handle correctly by ipvlan mode. Once the ipvlan interface is created and ip address assign to it and if we restart interface-config or database (new PR) service Linux Kernel gives error "Error: Address already assigned to an ipvlan device." based on this:https://github.com/torvalds/linux/blob/master/drivers/net/ipvlan/ipvlan_main.c#L978Reason being if we do not do cleanup of ip address assignment (need to be unique for IPVLAN) it remains in Kernel Database and never goes to free pool even though namespace is deleted. Solution: Considering this hard dependency of unique ip macvlan mode is better for us and since everything is managed by Linux Kernel and no dependency for on user configured IP address. Issue3: Namespace database Service do not check reachability to Supervisor Redis Chassis Server. Currently there is no explicit check as we never do Redis PING from namespace to Supervisor Redis Chassis Server. With this check it's possible we will start database and all other docker even though there is no connectivity and will hit the error/failure late in cycle Solution: Added explicit PING from namespace that will check this reachability. Issue 4:flushdb give exception when trying to accces Chassis Server DB over Unix Sokcet. Solution: Handle gracefully via try..except and log the message.
2022-05-24 18:54:12 -05:00
until [[ ($(docker exec -i database$DEV pgrep -x -c supervisord) -gt 0) && ($($SONIC_DB_CLI PING | grep -c PONG) -gt 0) &&
($(docker exec -i database$DEV sonic-db-cli PING | grep -c PONG) -gt 0) ]]; do
sleep 1;
done
if [[ ("$BOOT_TYPE" == "warm" || "$BOOT_TYPE" == "fastfast") && -f $WARM_DIR/dump.rdb ]]; then
# retain the dump file from last boot for debugging purposes
mv $WARM_DIR/dump.rdb $WARM_DIR/dump.rdb.old
else
# If there is a config_db.json dump file, load it.
if [ -r /etc/sonic/config_db$DEV.json ]; then
# For multi-asic, all /var/run/redis$DEV/sonic-db/database_config.json need to ready
# for loading config with --write-to-db
waitForAllInstanceDatabaseConfigJsonFilesReady
if [ -r /etc/sonic/init_cfg.json ]; then
$SONIC_CFGGEN -j /etc/sonic/init_cfg.json -j /etc/sonic/config_db$DEV.json --write-to-db
else
$SONIC_CFGGEN -j /etc/sonic/config_db$DEV.json --write-to-db
fi
fi
if [[ "$BOOT_TYPE" == "fast" ]]; then
# set the key to expire in 3 minutes
$SONIC_DB_CLI STATE_DB SET "FAST_REBOOT|system" "1" "EX" "180"
fi
$SONIC_DB_CLI CONFIG_DB SET "CONFIG_DB_INITIALIZED" "1"
fi
if [[ -x /usr/local/bin/db_migrator.py ]]; then
# Migrate the DB to the latest schema version if needed
if [ -z "$DEV" ]; then
/usr/local/bin/db_migrator.py -o migrate
fi
fi
# Add redis UDS to the redis group and give read/write access to the group
REDIS_SOCK="/var/run/redis${DEV}/redis.sock"
else
until [[ ($(docker exec -i ${DOCKERNAME} pgrep -x -c supervisord) -gt 0) &&
($(docker exec -i ${DOCKERNAME} $SONIC_DB_CLI CHASSIS_APP_DB PING | grep -c True) -gt 0) ]]; do
sleep 1
done
[chassis] Fix issues regarding database service failure handling and mid-plane connectivity for namespace. (#10500) What/Why I did: Issue1: By setting up of ipvlan interface in interface-config.sh we are not tolerant to failures. Reason being interface-config.service is one-shot and do not have restart capability. Scenario: For example if let's say database service goes in fail state then interface-services also gets failed because of dependency check but later database service gets restart but interface service will remain in stuck state and the ipvlan interface nevers get created. Solution: Moved all the logic in database service from interface-config service which looks more align logically also since the namespace is created here and all the network setting (sysctl) are happening here.With this if database starts we recreate the interface. Issue 2: Use of IPVLAN vs MACVLAN Currently we are using ipvlan mode. However above failure scenario is not handle correctly by ipvlan mode. Once the ipvlan interface is created and ip address assign to it and if we restart interface-config or database (new PR) service Linux Kernel gives error "Error: Address already assigned to an ipvlan device." based on this:https://github.com/torvalds/linux/blob/master/drivers/net/ipvlan/ipvlan_main.c#L978Reason being if we do not do cleanup of ip address assignment (need to be unique for IPVLAN) it remains in Kernel Database and never goes to free pool even though namespace is deleted. Solution: Considering this hard dependency of unique ip macvlan mode is better for us and since everything is managed by Linux Kernel and no dependency for on user configured IP address. Issue3: Namespace database Service do not check reachability to Supervisor Redis Chassis Server. Currently there is no explicit check as we never do Redis PING from namespace to Supervisor Redis Chassis Server. With this check it's possible we will start database and all other docker even though there is no connectivity and will hit the error/failure late in cycle Solution: Added explicit PING from namespace that will check this reachability. Issue 4:flushdb give exception when trying to accces Chassis Server DB over Unix Sokcet. Solution: Handle gracefully via try..except and log the message.
2022-05-24 18:54:12 -05:00
if [[ -n "$lag_id_start" && -n "$lag_id_end" ]]; then
setPlatformLagIdBoundaries
fi
REDIS_SOCK="/var/run/redis-chassis/redis_chassis.sock"
fi
chgrp -f redis $REDIS_SOCK && chmod -f 0760 $REDIS_SOCK
{%- elif docker_container_name == "swss" %}
docker exec swss$DEV rm -f /ready # remove cruft
if [[ "$BOOT_TYPE" == "fast" ]] && [[ -d /host/fast-reboot ]]; then
test -e /host/fast-reboot/fdb.json && docker cp /host/fast-reboot/fdb.json swss$DEV:/
test -e /host/fast-reboot/arp.json && docker cp /host/fast-reboot/arp.json swss$DEV:/
test -e /host/fast-reboot/default_routes.json && docker cp /host/fast-reboot/default_routes.json swss$DEV:/
test -e /host/fast-reboot/media_config.json && docker cp /host/fast-reboot/media_config.json swss$DEV:/
rm -fr /host/fast-reboot
fi
docker exec swss$DEV touch /ready # signal swssconfig.sh to go
{%- elif docker_container_name == "pmon" %}
DEVPATH="/usr/share/sonic/device"
REBOOT="platform_reboot"
PSENSOR="/usr/local/bin/platform_sensors.py"
if [ -d ${DEVPATH}/${PLATFORM} ] && [ -f $PSENSOR ]; then
exist=`docker exec -i pmon ls /usr/bin/platform_sensors.py "$@" 2>/dev/null`
if [ -z "$exist" ]; then
docker cp $PSENSOR pmon:/usr/bin/
fi
fi
{%- else %}
: # nothing
{%- endif %}
}
start() {
# Obtain boot type from kernel arguments
BOOT_TYPE=`getBootType`
# Obtain our platform as we will mount directories with these names in each docker
PLATFORM=${PLATFORM:-`$SONIC_CFGGEN -H -v DEVICE_METADATA.localhost.platform`}
# Parse the device specific asic conf file, if it exists
ASIC_CONF=/usr/share/sonic/device/$PLATFORM/asic.conf
if [ -f "$ASIC_CONF" ]; then
source $ASIC_CONF
fi
PLATFORM_ENV_CONF=/usr/share/sonic/device/$PLATFORM/platform_env.conf
if [ -f "$PLATFORM_ENV_CONF" ]; then
source $PLATFORM_ENV_CONF
fi
{%- if docker_container_name == "gbsyncd" %}
GBSYNCD_CONF=/usr/share/sonic/device/$PLATFORM/gbsyncd.ini
GBSYNCD_PLATFORM=gbsyncd-vs
if [ -f "$GBSYNCD_CONF" ]; then
while IFS="=" read -r key value; do
case "$key" in
platform)
GBSYNCD_PLATFORM="$value"
;;
esac
done < "$GBSYNCD_CONF"
fi
{%- endif %}
{%- if docker_container_name == "database" %}
# Don't mount HWSKU in {{docker_container_name}} container.
HWSKU=""
MOUNTPATH=""
{%- else %}
# Obtain our HWSKU as we will mount directories with these names in each docker
HWSKU=${HWSKU:-`$SONIC_CFGGEN -d -v 'DEVICE_METADATA["localhost"]["hwsku"]'`}
MOUNTPATH="/usr/share/sonic/device/$PLATFORM/$HWSKU"
if [ "$DEV" ]; then
MOUNTPATH="$MOUNTPATH/$DEV"
fi
{%- endif %}
DOCKERCHECK=`docker inspect --type container ${DOCKERNAME} 2>/dev/null`
if [ "$?" -eq "0" ]; then
{%- if docker_container_name == "database" %}
DOCKERMOUNT=""
{%- else %}
DOCKERMOUNT=`getMountPoint "$DOCKERCHECK"`
{%- endif %}
if [ x"$DOCKERMOUNT" == x"$MOUNTPATH" ]; then
preStartAction
{%- if docker_container_name == "database" %}
echo "Starting existing ${DOCKERNAME} container"
docker start ${DOCKERNAME}
{%- else %}
echo "Starting existing ${DOCKERNAME} container with HWSKU $HWSKU"
/usr/local/bin/container start ${DOCKERNAME}
{%- endif %}
postStartAction
exit $?
fi
# docker created with a different HWSKU, remove and recreate
echo "Removing obsolete ${DOCKERNAME} container with HWSKU $DOCKERMOUNT"
docker rm -f ${DOCKERNAME}
fi
{%- if docker_container_name == "database" %}
echo "Creating new ${DOCKERNAME} container"
if [ "$DATABASE_TYPE" != "chassisdb" ]; then
if [ -z "$DEV" ]; then
# if database_global exists in old_config, use it; otherwise use the default one in new image
if [ -f /etc/sonic/old_config/database_global.json ]; then
echo "Use database_global.json from old system..."
mv /etc/sonic/old_config/database_global.json /etc/sonic/
fi
fi
# if database_config exists in old_config, use it; otherwise use the default one in new image
if [ -f /etc/sonic/old_config/database_config$DEV.json ]; then
echo "Use database_config.json from old system..."
mv /etc/sonic/old_config/database_config$DEV.json /etc/sonic/
fi
create multiple Redis DB instances based on CONFIG at /etc/sonic/database_config.json (#2182) this is the first step to moving different databases tables into different database instances in this PR, only handle multiple database instances creation based on user configuration at /etc/sonic/database_config.json we keep current method to create single database instance if no extra/new DATABASE configuration exist in database_config.json file. if user try to configure more db instances at database_config.json , we create those new db instances along with the original db instance existing today. The configuration is as below, later we can add more db related information if needed: { ... "DATABASE": { "redis-db-01" : { "port" : "6380", "database": ["APPL_DB", "STATE_DB"] }, "redis-db-02" : { "port" : "6381", "database":["ASIC_DB"] }, } ... } The detail description is at design doc at Azure/SONiC#271 The main idea is : when database.sh started, we check the configuration and generate corresponding scripts. rc.local service handle old_config copy when loading new images, there is no dependency between rc.local and database service today, for safety and make sure the copy operation are done before database try to read it, we make database service run after rc.local Then database docker started, we check the configuration and generate corresponding scripts/.conf in database docker as well. based on those conf, we create databases instances as required. at last, we ping_pong check database are up and continue 
Signed-off-by: Dong Zhang d.zhang@alibaba-inc.com
2019-08-28 13:15:10 -05:00
fi
{%- else %}
echo "Creating new ${DOCKERNAME} container with HWSKU $HWSKU"
{%- endif %}
[Dynamic buffer calc] Support dynamic buffer calculation (#6194) **- Why I did it** To support dynamic buffer calculation. This PR also depends on the following PRs for sub modules - [sonic-swss: [buffermgr/bufferorch] Support dynamic buffer calculation #1338](https://github.com/Azure/sonic-swss/pull/1338) - [sonic-swss-common: Dynamic buffer calculation #361](https://github.com/Azure/sonic-swss-common/pull/361) - [sonic-utilities: Support dynamic buffer calculation #973](https://github.com/Azure/sonic-utilities/pull/973) **- How I did it** 1. Introduce field `buffer_model` in `DEVICE_METADATA|localhost` to represent which buffer model is running in the system currently: - `dynamic` for the dynamic buffer calculation model - `traditional` for the traditional model in which the `pg_profile_lookup.ini` is used 2. Add the tables required for the feature: - ASIC_TABLE in platform/\<vendor\>/asic_table.j2 - PERIPHERAL_TABLE in platform/\<vendor\>/peripheral_table.j2 - PORT_PERIPHERAL_TABLE on a per-platform basis in device/\<vendor\>/\<platform\>/port_peripheral_config.j2 for each platform with gearbox installed. - DEFAULT_LOSSLESS_BUFFER_PARAMETER and LOSSLESS_TRAFFIC_PATTERN in files/build_templates/buffers_config.j2 - Add lossless PGs (3-4) for each port in files/build_templates/buffers_config.j2 3. Copy the newly introduced j2 files into the image and rendering them when the system starts 4. Update the CLI options for buffermgrd so that it can start with dynamic mode 5. Fetches the ASIC vendor name in orchagent: - fetch the vendor name when creates the docker and pass it as a docker environment variable - `buffermgrd` can use this passed-in variable 6. Clear buffer related tables from STATE_DB when swss docker starts 7. Update the src/sonic-config-engine/tests/sample_output/buffers-dell6100.json according to the buffer_config.j2 8. Remove buffer pool sizes for ingress pools and egress_lossy_pool Update the buffer settings for dynamic buffer calculation
2020-12-13 13:35:39 -06:00
{%- if docker_container_name == "swss" %}
# Generate the asic_table.json and peripheral_table.json
if [ ! -f /etc/sonic/asic_table.json ] && [ -f /usr/share/sonic/templates/asic_table.j2 ]; then
sonic-cfggen -d -t /usr/share/sonic/templates/asic_table.j2 > /etc/sonic/asic_table.json
fi
if [ ! -f /etc/sonic/peripheral_table.json ] && [ -f /usr/share/sonic/device/$PLATFORM/port_peripheral_config.j2 ]; then
sonic-cfggen -d -t /usr/share/sonic/device/$PLATFORM/port_peripheral_config.j2 > /etc/sonic/peripheral_table.json
fi
if [ ! -f /etc/sonic/zero_profiles.json ] && [ -f /usr/share/sonic/templates/zero_profiles.j2 ]; then
sonic-cfggen -d -t /usr/share/sonic/device/$PLATFORM/zero_profiles.j2 > /etc/sonic/zero_profiles.json
fi
{%- if enable_asan == "y" %}
mkdir -p /var/log/asan
{%- endif %}
[Dynamic buffer calc] Support dynamic buffer calculation (#6194) **- Why I did it** To support dynamic buffer calculation. This PR also depends on the following PRs for sub modules - [sonic-swss: [buffermgr/bufferorch] Support dynamic buffer calculation #1338](https://github.com/Azure/sonic-swss/pull/1338) - [sonic-swss-common: Dynamic buffer calculation #361](https://github.com/Azure/sonic-swss-common/pull/361) - [sonic-utilities: Support dynamic buffer calculation #973](https://github.com/Azure/sonic-utilities/pull/973) **- How I did it** 1. Introduce field `buffer_model` in `DEVICE_METADATA|localhost` to represent which buffer model is running in the system currently: - `dynamic` for the dynamic buffer calculation model - `traditional` for the traditional model in which the `pg_profile_lookup.ini` is used 2. Add the tables required for the feature: - ASIC_TABLE in platform/\<vendor\>/asic_table.j2 - PERIPHERAL_TABLE in platform/\<vendor\>/peripheral_table.j2 - PORT_PERIPHERAL_TABLE on a per-platform basis in device/\<vendor\>/\<platform\>/port_peripheral_config.j2 for each platform with gearbox installed. - DEFAULT_LOSSLESS_BUFFER_PARAMETER and LOSSLESS_TRAFFIC_PATTERN in files/build_templates/buffers_config.j2 - Add lossless PGs (3-4) for each port in files/build_templates/buffers_config.j2 3. Copy the newly introduced j2 files into the image and rendering them when the system starts 4. Update the CLI options for buffermgrd so that it can start with dynamic mode 5. Fetches the ASIC vendor name in orchagent: - fetch the vendor name when creates the docker and pass it as a docker environment variable - `buffermgrd` can use this passed-in variable 6. Clear buffer related tables from STATE_DB when swss docker starts 7. Update the src/sonic-config-engine/tests/sample_output/buffers-dell6100.json according to the buffer_config.j2 8. Remove buffer pool sizes for ingress pools and egress_lossy_pool Update the buffer settings for dynamic buffer calculation
2020-12-13 13:35:39 -06:00
{%- endif %}
# In Multi ASIC platforms the global database config file database_global.json will exist.
# Parse the file and get the include path for the database_config.json files used in
# various namesapces. The database_config paths are relative to the DIR of SONIC_DB_GLOBAL_JSON.
SONIC_DB_GLOBAL_JSON="/var/run/redis/sonic-db/database_global.json"
if [ -f "$SONIC_DB_GLOBAL_JSON" ]; then
# TODO Create a separate python script with the below logic and invoke it here.
redis_dir_list=`/usr/bin/python -c "import sys; import os; import json; f=open(sys.argv[1]); \
global_db_dir = os.path.dirname(sys.argv[1]); data=json.load(f); \
print(\" \".join([os.path.normpath(global_db_dir+'/'+elem['include']).partition('sonic-db')[0]\
for elem in data['INCLUDES'] if 'namespace' in elem])); f.close()" $SONIC_DB_GLOBAL_JSON`
fi
{%- if docker_container_name == "database" %}
start_chassis_db=0
chassis_db_address=""
chassisdb_config="/usr/share/sonic/device/$PLATFORM/chassisdb.conf"
[ -f $chassisdb_config ] && source $chassisdb_config
DB_OPT=" -v /var/run/redis-chassis:/var/run/redis-chassis:ro "
if [[ "$start_chassis_db" != "1" ]] && [[ -z "$chassis_db_address" ]]; then
DB_OPT=""
else
DB_OPT=$DB_OPT" --add-host=redis_chassis.server:$chassis_db_address "
fi
{%- endif %}
if [ -z "$DEV" ]; then
NET="host"
# For Multi-ASIC platform we have to mount the redis paths for database instances running in different
# namespaces, into the single instance dockers like snmp, pmon on linux host. These global dockers
# will need to get/set tables from databases in different namespaces.
# /var/run/redis0 ---> mounted as --> /var/run/redis0
# /var/run/redis1 ---> mounted as --> /var/run/redis1 .. etc
# The below logic extracts the base DIR's where database_config.json's for various namespaces exist.
# redis_dir_list is a string of form "/var/run/redis0/ /var/run/redis1/ /var/run/redis2/"
{%- if docker_container_name != "database" %}
if [ -n "$redis_dir_list" ]; then
for redis_dir in $redis_dir_list
do
REDIS_MNT=$REDIS_MNT" -v $redis_dir:$redis_dir:rw "
done
fi
{%- else %}
if [ "$DATABASE_TYPE" == "chassisdb" ]; then
DB_OPT=${DB_OPT/redis-chassis:ro/redis-chassis:rw}
DB_OPT=$DB_OPT" -v /var/run/redis-chassis:/var/run/redis:rw "
DB_OPT=$DB_OPT" --env DATABASE_TYPE=$DATABASE_TYPE"
else
DB_OPT=$DB_OPT" -v /var/run/redis$DEV:/var/run/redis:rw "
fi
{%- endif %}
else
# This part of code is applicable for Multi-ASIC platforms. Here we mount the namespace specific
# redis directory into the docker running in that namespace. Below eg: is for namespace "asic1"
# /var/run/redis1 ---> mounted as --> /var/run/redis1
# redis_dir_list is a string of form "/var/run/redis0/ /var/run/redis1/ /var/run/redis2/"
if [ -n "$redis_dir_list" ]; then
id=`expr $DEV + 1`
redis_dir=`echo $redis_dir_list | cut -d " " -f $id`
REDIS_MNT=" -v $redis_dir:$redis_dir:rw "
fi
{%- if docker_container_name == "database" %}
NET="bridge"
DB_OPT=$DB_OPT" -v /var/run/redis$DEV:/var/run/redis:rw "
{%- else %}
NET="container:database$DEV"
DB_OPT=""
{%- endif %}
fi
{%- if docker_container_name == "bgp" %}
if [ "$DEV" ]; then
if [ ! -d "/etc/sonic/frr/$DEV" ]; then
mkdir /etc/sonic/frr/$DEV
cp -r /etc/sonic/frr/*.conf /etc/sonic/frr/$DEV
fi
fi
{%- endif %}
{%- if sonic_asic_platform == "mellanox" %}
# TODO: Mellanox will remove the --tmpfs exception after SDK socket path changed in new SDK version
{%- endif %}
docker create {{docker_image_run_opt}} \
--net=$NET \
-e RUNTIME_OWNER=local \
--uts=host \{# W/A: this should be set per-docker, for those dockers which really need host's UTS namespace #}
{%- if install_debug_image == "y" %}
-v /src:/src:ro -v /debug:/debug:rw \
{%- endif %}
{%- if '--log-driver=json-file' in docker_image_run_opt or '--log-driver' not in docker_image_run_opt %}
--log-opt max-size=2M --log-opt max-file=5 \
{%- endif %}
{%- if sonic_asic_platform == "mellanox" %}
{%- if docker_container_name == "syncd" %}
-v /var/log/mellanox:/var/log/mellanox:rw \
-v mlnx_sdk_socket:/var/run/sx_sdk \
-v mlnx_sdk_ready:/tmp \
-v /dev/shm:/dev/shm:rw \
-e SX_API_SOCKET_FILE=/var/run/sx_sdk/sx_api.sock \
{%- elif docker_container_name == "pmon" %}
-v /var/run/hw-management:/var/run/hw-management:rw \
-v mlnx_sdk_socket:/var/run/sx_sdk \
-v mlnx_sdk_ready:/tmp \
-e SX_API_SOCKET_FILE=/var/run/sx_sdk/sx_api.sock \
-v /dev/shm:/dev/shm:rw \
{%- else %}
{%- if mount_default_tmpfs|default("n") == "y" %}
--tmpfs /tmp \
{%- endif %}
{%- endif %}
{%- endif %}
{%- if sonic_asic_platform == "broadcom" %}
{%- if docker_container_name == "syncd" %}
--shm-size=${SYNCD_SHM_SIZE:-64m} \
-v /var/run/docker-syncd$DEV:/var/run/sswsyncd \
{%- endif %}
{%- endif %}
{%- if docker_container_name == "pmon" %}
-v /usr/share/sonic/firmware:/usr/share/sonic/firmware:rw \
{%- endif %}
[Dynamic buffer calc] Support dynamic buffer calculation (#6194) **- Why I did it** To support dynamic buffer calculation. This PR also depends on the following PRs for sub modules - [sonic-swss: [buffermgr/bufferorch] Support dynamic buffer calculation #1338](https://github.com/Azure/sonic-swss/pull/1338) - [sonic-swss-common: Dynamic buffer calculation #361](https://github.com/Azure/sonic-swss-common/pull/361) - [sonic-utilities: Support dynamic buffer calculation #973](https://github.com/Azure/sonic-utilities/pull/973) **- How I did it** 1. Introduce field `buffer_model` in `DEVICE_METADATA|localhost` to represent which buffer model is running in the system currently: - `dynamic` for the dynamic buffer calculation model - `traditional` for the traditional model in which the `pg_profile_lookup.ini` is used 2. Add the tables required for the feature: - ASIC_TABLE in platform/\<vendor\>/asic_table.j2 - PERIPHERAL_TABLE in platform/\<vendor\>/peripheral_table.j2 - PORT_PERIPHERAL_TABLE on a per-platform basis in device/\<vendor\>/\<platform\>/port_peripheral_config.j2 for each platform with gearbox installed. - DEFAULT_LOSSLESS_BUFFER_PARAMETER and LOSSLESS_TRAFFIC_PATTERN in files/build_templates/buffers_config.j2 - Add lossless PGs (3-4) for each port in files/build_templates/buffers_config.j2 3. Copy the newly introduced j2 files into the image and rendering them when the system starts 4. Update the CLI options for buffermgrd so that it can start with dynamic mode 5. Fetches the ASIC vendor name in orchagent: - fetch the vendor name when creates the docker and pass it as a docker environment variable - `buffermgrd` can use this passed-in variable 6. Clear buffer related tables from STATE_DB when swss docker starts 7. Update the src/sonic-config-engine/tests/sample_output/buffers-dell6100.json according to the buffer_config.j2 8. Remove buffer pool sizes for ingress pools and egress_lossy_pool Update the buffer settings for dynamic buffer calculation
2020-12-13 13:35:39 -06:00
{%- if docker_container_name == "swss" %}
-e ASIC_VENDOR={{ sonic_asic_platform }} \
{%- endif -%}
{%- if docker_container_name in ["swss", "syncd"] and enable_asan == "y" %}
-v /var/log/asan/:/var/log/asan \
[Dynamic buffer calc] Support dynamic buffer calculation (#6194) **- Why I did it** To support dynamic buffer calculation. This PR also depends on the following PRs for sub modules - [sonic-swss: [buffermgr/bufferorch] Support dynamic buffer calculation #1338](https://github.com/Azure/sonic-swss/pull/1338) - [sonic-swss-common: Dynamic buffer calculation #361](https://github.com/Azure/sonic-swss-common/pull/361) - [sonic-utilities: Support dynamic buffer calculation #973](https://github.com/Azure/sonic-utilities/pull/973) **- How I did it** 1. Introduce field `buffer_model` in `DEVICE_METADATA|localhost` to represent which buffer model is running in the system currently: - `dynamic` for the dynamic buffer calculation model - `traditional` for the traditional model in which the `pg_profile_lookup.ini` is used 2. Add the tables required for the feature: - ASIC_TABLE in platform/\<vendor\>/asic_table.j2 - PERIPHERAL_TABLE in platform/\<vendor\>/peripheral_table.j2 - PORT_PERIPHERAL_TABLE on a per-platform basis in device/\<vendor\>/\<platform\>/port_peripheral_config.j2 for each platform with gearbox installed. - DEFAULT_LOSSLESS_BUFFER_PARAMETER and LOSSLESS_TRAFFIC_PATTERN in files/build_templates/buffers_config.j2 - Add lossless PGs (3-4) for each port in files/build_templates/buffers_config.j2 3. Copy the newly introduced j2 files into the image and rendering them when the system starts 4. Update the CLI options for buffermgrd so that it can start with dynamic mode 5. Fetches the ASIC vendor name in orchagent: - fetch the vendor name when creates the docker and pass it as a docker environment variable - `buffermgrd` can use this passed-in variable 6. Clear buffer related tables from STATE_DB when swss docker starts 7. Update the src/sonic-config-engine/tests/sample_output/buffers-dell6100.json according to the buffer_config.j2 8. Remove buffer pool sizes for ingress pools and egress_lossy_pool Update the buffer settings for dynamic buffer calculation
2020-12-13 13:35:39 -06:00
{%- endif -%}
{%- if docker_container_name == "bgp" %}
-v /etc/sonic/frr/$DEV:/etc/frr:rw \
{%- endif %}
{%- if docker_container_name == "database" %}
$DB_OPT \
{%- else %}
-v /var/run/redis$DEV:/var/run/redis:rw \
-v /var/run/redis-chassis:/var/run/redis-chassis:ro \
-v /usr/share/sonic/device/$PLATFORM/$HWSKU/$DEV:/usr/share/sonic/hwsku:ro \
{%- endif %}
$REDIS_MNT \
-v /usr/share/sonic/device/$PLATFORM:/usr/share/sonic/platform:ro \
{%- if sonic_asic_platform != "mellanox" %}
{%- if mount_default_tmpfs|default("n") == "y" %}
--tmpfs /tmp \
{%- endif %}
{%- endif %}
{%- if mount_default_tmpfs|default("n") == "y" %}
--tmpfs /var/tmp \
{%- endif %}
--env "NAMESPACE_ID"="$DEV" \
--env "NAMESPACE_PREFIX"="$NAMESPACE_PREFIX" \
--env "NAMESPACE_COUNT"=$NUM_ASIC \
--name=$DOCKERNAME \
{%- if docker_container_name == "gbsyncd" %}
-v /var/run/docker-syncd$DEV:/var/run/sswsyncd \
"docker-$GBSYNCD_PLATFORM":latest \
{%- elif docker_image_name is defined %}
{{docker_image_name}}:latest \
{%- else %}
{{docker_image_id}} \
{%- endif %}
|| {
echo "Failed to docker run" >&1
exit 4
}
preStartAction
{%- if docker_container_name == "database" %}
docker start $DOCKERNAME
{%- else %}
/usr/local/bin/container start ${DOCKERNAME}
{%- endif %}
postStartAction
}
wait() {
{%- if docker_container_name == "database" %}
docker wait $DOCKERNAME
{%- else %}
/usr/local/bin/container wait $DOCKERNAME
{%- endif %}
}
stop() {
{%- if docker_container_name == "database" %}
docker stop $DOCKERNAME
if [ "$DEV" ]; then
ip netns delete "$NET_NS"
fi
{%- elif docker_container_name == "teamd" %}
# Longer timeout of 60 sec to wait for Portchannels to be cleaned.
/usr/local/bin/container stop -t 60 $DOCKERNAME
{%- elif docker_container_name in ["swss", "syncd"] and enable_asan == "y" %}
/usr/local/bin/container stop -t 60 $DOCKERNAME
{%- else %}
/usr/local/bin/container stop $DOCKERNAME
{%- endif %}
}
DOCKERNAME={{docker_container_name}}
OP=$1
DEV=$2 # namespace/device number to operate on
{%- if docker_container_name == "database" %}
if [ "$DEV" == "chassisdb" ]; then
DATABASE_TYPE="chassisdb"
DOCKERNAME=$DOCKERNAME"-chassis"
unset DEV
fi
{%- endif %}
NAMESPACE_PREFIX="asic"
DOCKERNAME=$DOCKERNAME$DEV
if [ "$DEV" ]; then
NET_NS="$NAMESPACE_PREFIX$DEV" #name of the network namespace
SONIC_CFGGEN="sonic-cfggen -n $NET_NS"
SONIC_DB_CLI="sonic-db-cli -n $NET_NS"
else
NET_NS=""
SONIC_CFGGEN="sonic-cfggen"
SONIC_DB_CLI="sonic-db-cli"
fi
# read SONiC immutable variables
[ -f /etc/sonic/sonic-environment ] && . /etc/sonic/sonic-environment
case "$1" in
start|wait|stop)
$1
;;
*)
echo "Usage: $0 {start namespace(optional)|wait namespace(optional)|stop namespace(optional)}"
exit 1
;;
esac