2018-09-24 18:35:01 -05:00
|
|
|
#!/bin/bash
|
|
|
|
|
|
|
|
SERVICE="syncd"
|
|
|
|
PEER="swss"
|
|
|
|
DEBUGLOG="/tmp/swss-syncd-debug.log"
|
|
|
|
LOCKFILE="/tmp/swss-syncd-lock"
|
|
|
|
|
|
|
|
function debug()
|
|
|
|
{
|
2019-02-10 13:56:31 -06:00
|
|
|
/usr/bin/logger $1
|
2018-09-24 18:35:01 -05:00
|
|
|
/bin/echo `date` "- $1" >> ${DEBUGLOG}
|
|
|
|
}
|
|
|
|
|
|
|
|
function lock_service_state_change()
|
|
|
|
{
|
|
|
|
debug "Locking ${LOCKFILE} from ${SERVICE} service"
|
|
|
|
|
|
|
|
exec {LOCKFD}>${LOCKFILE}
|
|
|
|
/usr/bin/flock -x ${LOCKFD}
|
|
|
|
trap "/usr/bin/flock -u ${LOCKFD}" 0 2 3 15
|
|
|
|
|
|
|
|
debug "Locked ${LOCKFILE} (${LOCKFD}) from ${SERVICE} service"
|
|
|
|
}
|
|
|
|
|
|
|
|
function unlock_service_state_change()
|
|
|
|
{
|
|
|
|
debug "Unlocking ${LOCKFILE} (${LOCKFD}) from ${SERVICE} service"
|
|
|
|
/usr/bin/flock -u ${LOCKFD}
|
|
|
|
}
|
|
|
|
|
|
|
|
function check_warm_boot()
|
|
|
|
{
|
2019-02-19 19:06:56 -06:00
|
|
|
SYSTEM_WARM_START=`/usr/bin/redis-cli -n 6 hget "WARM_RESTART_ENABLE_TABLE|system" enable`
|
|
|
|
SERVICE_WARM_START=`/usr/bin/redis-cli -n 6 hget "WARM_RESTART_ENABLE_TABLE|${SERVICE}" enable`
|
2018-09-24 18:35:01 -05:00
|
|
|
# SYSTEM_WARM_START could be empty, always make WARM_BOOT meaningful.
|
2018-10-16 13:20:39 -05:00
|
|
|
if [[ x"$SYSTEM_WARM_START" == x"true" ]] || [[ x"$SERVICE_WARM_START" == x"true" ]]; then
|
2018-09-24 18:35:01 -05:00
|
|
|
WARM_BOOT="true"
|
|
|
|
else
|
|
|
|
WARM_BOOT="false"
|
|
|
|
fi
|
|
|
|
}
|
|
|
|
|
|
|
|
function wait_for_database_service()
|
|
|
|
{
|
|
|
|
# Wait for redis server start before database clean
|
|
|
|
until [[ $(/usr/bin/docker exec database redis-cli ping | grep -c PONG) -gt 0 ]];
|
|
|
|
do sleep 1;
|
|
|
|
done
|
|
|
|
|
|
|
|
# Wait for configDB initialization
|
|
|
|
until [[ $(/usr/bin/docker exec database redis-cli -n 4 GET "CONFIG_DB_INITIALIZED") ]];
|
|
|
|
do sleep 1;
|
|
|
|
done
|
|
|
|
}
|
|
|
|
|
2018-12-15 13:36:12 -06:00
|
|
|
function getBootType()
|
|
|
|
{
|
2019-06-26 14:46:58 -05:00
|
|
|
# same code snippet in files/build_templates/docker_image_ctl.j2
|
|
|
|
case "$(cat /proc/cmdline)" in
|
|
|
|
*SONIC_BOOT_TYPE=warm*)
|
2018-12-15 13:36:12 -06:00
|
|
|
TYPE='warm'
|
|
|
|
;;
|
2019-06-26 14:46:58 -05:00
|
|
|
*SONIC_BOOT_TYPE=fastfast*)
|
2018-12-15 13:36:12 -06:00
|
|
|
TYPE='fastfast'
|
|
|
|
;;
|
2019-06-26 14:46:58 -05:00
|
|
|
*SONIC_BOOT_TYPE=fast*|*fast-reboot*)
|
2019-08-19 18:05:20 -05:00
|
|
|
TYPE=$(awk '{ if ($1 <= 180) print "fast"; else print "cold" }' /proc/uptime)
|
2018-12-15 13:36:12 -06:00
|
|
|
;;
|
|
|
|
*)
|
|
|
|
TYPE='cold'
|
|
|
|
esac
|
|
|
|
echo "${TYPE}"
|
|
|
|
}
|
|
|
|
|
2018-09-24 18:35:01 -05:00
|
|
|
start() {
|
|
|
|
debug "Starting ${SERVICE} service..."
|
|
|
|
|
|
|
|
lock_service_state_change
|
|
|
|
|
2018-10-01 21:01:04 -05:00
|
|
|
mkdir -p /host/warmboot
|
|
|
|
|
2018-09-24 18:35:01 -05:00
|
|
|
wait_for_database_service
|
|
|
|
check_warm_boot
|
|
|
|
|
|
|
|
debug "Warm boot flag: ${SERVICE} ${WARM_BOOT}."
|
|
|
|
|
2018-11-21 19:56:19 -06:00
|
|
|
if [[ x"$WARM_BOOT" == x"true" ]]; then
|
|
|
|
# Leave a mark for syncd scripts running inside docker.
|
|
|
|
touch /host/warmboot/warm-starting
|
|
|
|
else
|
|
|
|
rm -f /host/warmboot/warm-starting
|
2018-09-24 18:35:01 -05:00
|
|
|
fi
|
|
|
|
|
2018-12-15 13:36:12 -06:00
|
|
|
# platform specific tasks
|
2019-01-10 16:09:04 -06:00
|
|
|
|
|
|
|
# start mellanox drivers regardless of
|
|
|
|
# boot type
|
2019-03-06 20:51:46 -06:00
|
|
|
if [[ x"$sonic_asic_platform" == x"mellanox" ]]; then
|
2018-12-15 13:36:12 -06:00
|
|
|
BOOT_TYPE=`getBootType`
|
|
|
|
if [[ x"$WARM_BOOT" == x"true" || x"$BOOT_TYPE" == x"fast" ]]; then
|
|
|
|
export FAST_BOOT=1
|
|
|
|
fi
|
2019-03-06 20:51:46 -06:00
|
|
|
|
|
|
|
if [[ x"$WARM_BOOT" != x"true" ]]; then
|
2019-05-29 01:57:29 -05:00
|
|
|
if [[ x"$(/bin/systemctl is-active pmon)" == x"active" ]]; then
|
|
|
|
/bin/systemctl stop pmon
|
[Mellanox] Stop pmon ahead of syncd (#3505)
Issue Overview
shutdown flow
For any shutdown flow, which means all dockers are stopped in order, pmon docker stops after syncd docker has stopped, causing pmon docker fail to release sx_core resources and leaving sx_core in a bad state. The related logs are like the following:
INFO syncd.sh[23597]: modprobe: FATAL: Module sx_core is in use.
INFO syncd.sh[23597]: Unloading sx_core[FAILED]
INFO syncd.sh[23597]: rmmod: ERROR: Module sx_core is in use
config reload & service swss.restart
In the flows like "config reload" and "service swss restart", the failure cause further consequences:
sx_core initialization error with error message like "sx_core: create EMAD sdq 0 failed. err: -16"
syncd fails to execute the create switch api with error message "syncd_main: Runtime error: :- processEvent: failed to execute api: create, key: SAI_OBJECT_TYPE_SWITCH:oid:0x21000000000000, status: SAI_STATUS_FAILURE"
swss fails to call SAI API "SAI_SWITCH_ATTR_INIT_SWITCH", which causes orchagent to restart. This will introduce an extra 1 or 2 minutes for the system to be available, failing related test cases.
reboot, warm-reboot & fast-reboot
In the reboot flows including "reboot", "fast-reboot" and "warm-reboot" this failure doesn't have further negative effects since the system has already rebooted. In addition, "warm-reboot" requires the system to be shutdown as soon as possible to meet the GR time restriction of both BGP and LACP. "fast-reboot" also requires to meet the GR time restriction of BGP which is longer than LACP. In this sense, any unnecessary steps should be avoided. It's better to keep those flows untouched.
summary
To summarize, we have to come up with a way to ensure:
shutdown pmon docker ahead of syncd for "config reload" or "service swss restart" flow;
don't shutdown pmon docker ahead of syncd for "fast-reboot" or "warm-reboot" flow in order to save time.
for "reboot" flow, either order is acceptable.
Solution
To solve the issue, pmon shoud be stopped ahead of syncd stopped for all flows except for the warm-reboot.
- How I did it
To stop pmon ahead of syncd stopped. This is done in /usr/local/bin/syncd.sh::stop() and for all shutdown sequence.
Now pmon stops ahead of syncd so there must be a way in which pmon can start after syncd started. Another point that should be taken consideration is that pmon starting should be deferred so that services which have the logic of graceful restart in fast-reboot and warm-reboot have sufficient CPU cycles to meet their deadline.
This is done by add "syncd.service" as "After" to pmon.service and startin /usr/local/bin/syncd.sh::wait()
To start pmon automatically after syncd started.
2019-09-27 03:15:46 -05:00
|
|
|
debug "pmon is active while syncd starting, stop it first"
|
2019-05-29 01:57:29 -05:00
|
|
|
fi
|
2019-03-06 20:51:46 -06:00
|
|
|
fi
|
|
|
|
|
|
|
|
if [[ x"$BOOT_TYPE" == x"fast" ]]; then
|
|
|
|
/usr/bin/hw-management.sh chipupdis
|
|
|
|
fi
|
|
|
|
|
2018-12-15 13:36:12 -06:00
|
|
|
/usr/bin/mst start
|
|
|
|
/usr/bin/mlnx-fw-upgrade.sh
|
|
|
|
/etc/init.d/sxdkernel start
|
|
|
|
fi
|
|
|
|
|
2019-01-10 16:09:04 -06:00
|
|
|
if [[ x"$WARM_BOOT" != x"true" ]]; then
|
|
|
|
if [ x$sonic_asic_platform == x'cavium' ]; then
|
|
|
|
/etc/init.d/xpnet.sh start
|
|
|
|
fi
|
|
|
|
fi
|
|
|
|
|
2018-09-24 18:35:01 -05:00
|
|
|
# start service docker
|
|
|
|
/usr/bin/${SERVICE}.sh start
|
|
|
|
debug "Started ${SERVICE} service..."
|
|
|
|
|
|
|
|
unlock_service_state_change
|
2019-03-02 17:28:34 -06:00
|
|
|
}
|
|
|
|
|
2019-03-08 12:59:41 -06:00
|
|
|
wait() {
|
[Mellanox] Stop pmon ahead of syncd (#3505)
Issue Overview
shutdown flow
For any shutdown flow, which means all dockers are stopped in order, pmon docker stops after syncd docker has stopped, causing pmon docker fail to release sx_core resources and leaving sx_core in a bad state. The related logs are like the following:
INFO syncd.sh[23597]: modprobe: FATAL: Module sx_core is in use.
INFO syncd.sh[23597]: Unloading sx_core[FAILED]
INFO syncd.sh[23597]: rmmod: ERROR: Module sx_core is in use
config reload & service swss.restart
In the flows like "config reload" and "service swss restart", the failure cause further consequences:
sx_core initialization error with error message like "sx_core: create EMAD sdq 0 failed. err: -16"
syncd fails to execute the create switch api with error message "syncd_main: Runtime error: :- processEvent: failed to execute api: create, key: SAI_OBJECT_TYPE_SWITCH:oid:0x21000000000000, status: SAI_STATUS_FAILURE"
swss fails to call SAI API "SAI_SWITCH_ATTR_INIT_SWITCH", which causes orchagent to restart. This will introduce an extra 1 or 2 minutes for the system to be available, failing related test cases.
reboot, warm-reboot & fast-reboot
In the reboot flows including "reboot", "fast-reboot" and "warm-reboot" this failure doesn't have further negative effects since the system has already rebooted. In addition, "warm-reboot" requires the system to be shutdown as soon as possible to meet the GR time restriction of both BGP and LACP. "fast-reboot" also requires to meet the GR time restriction of BGP which is longer than LACP. In this sense, any unnecessary steps should be avoided. It's better to keep those flows untouched.
summary
To summarize, we have to come up with a way to ensure:
shutdown pmon docker ahead of syncd for "config reload" or "service swss restart" flow;
don't shutdown pmon docker ahead of syncd for "fast-reboot" or "warm-reboot" flow in order to save time.
for "reboot" flow, either order is acceptable.
Solution
To solve the issue, pmon shoud be stopped ahead of syncd stopped for all flows except for the warm-reboot.
- How I did it
To stop pmon ahead of syncd stopped. This is done in /usr/local/bin/syncd.sh::stop() and for all shutdown sequence.
Now pmon stops ahead of syncd so there must be a way in which pmon can start after syncd started. Another point that should be taken consideration is that pmon starting should be deferred so that services which have the logic of graceful restart in fast-reboot and warm-reboot have sufficient CPU cycles to meet their deadline.
This is done by add "syncd.service" as "After" to pmon.service and startin /usr/local/bin/syncd.sh::wait()
To start pmon automatically after syncd started.
2019-09-27 03:15:46 -05:00
|
|
|
if [[ x"$sonic_asic_platform" == x"mellanox" ]]; then
|
|
|
|
debug "Starting pmon service..."
|
|
|
|
/bin/systemctl start pmon
|
|
|
|
debug "Started pmon service"
|
|
|
|
fi
|
2019-03-08 12:59:41 -06:00
|
|
|
/usr/bin/${SERVICE}.sh wait
|
2018-09-24 18:35:01 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
stop() {
|
|
|
|
debug "Stopping ${SERVICE} service..."
|
|
|
|
|
|
|
|
lock_service_state_change
|
|
|
|
check_warm_boot
|
|
|
|
debug "Warm boot flag: ${SERVICE} ${WARM_BOOT}."
|
|
|
|
|
2018-10-01 21:01:04 -05:00
|
|
|
if [[ x"$WARM_BOOT" == x"true" ]]; then
|
2018-11-15 17:47:33 -06:00
|
|
|
TYPE=warm
|
|
|
|
else
|
|
|
|
TYPE=cold
|
|
|
|
fi
|
2018-10-01 21:01:04 -05:00
|
|
|
|
[Mellanox] Stop pmon ahead of syncd (#3505)
Issue Overview
shutdown flow
For any shutdown flow, which means all dockers are stopped in order, pmon docker stops after syncd docker has stopped, causing pmon docker fail to release sx_core resources and leaving sx_core in a bad state. The related logs are like the following:
INFO syncd.sh[23597]: modprobe: FATAL: Module sx_core is in use.
INFO syncd.sh[23597]: Unloading sx_core[FAILED]
INFO syncd.sh[23597]: rmmod: ERROR: Module sx_core is in use
config reload & service swss.restart
In the flows like "config reload" and "service swss restart", the failure cause further consequences:
sx_core initialization error with error message like "sx_core: create EMAD sdq 0 failed. err: -16"
syncd fails to execute the create switch api with error message "syncd_main: Runtime error: :- processEvent: failed to execute api: create, key: SAI_OBJECT_TYPE_SWITCH:oid:0x21000000000000, status: SAI_STATUS_FAILURE"
swss fails to call SAI API "SAI_SWITCH_ATTR_INIT_SWITCH", which causes orchagent to restart. This will introduce an extra 1 or 2 minutes for the system to be available, failing related test cases.
reboot, warm-reboot & fast-reboot
In the reboot flows including "reboot", "fast-reboot" and "warm-reboot" this failure doesn't have further negative effects since the system has already rebooted. In addition, "warm-reboot" requires the system to be shutdown as soon as possible to meet the GR time restriction of both BGP and LACP. "fast-reboot" also requires to meet the GR time restriction of BGP which is longer than LACP. In this sense, any unnecessary steps should be avoided. It's better to keep those flows untouched.
summary
To summarize, we have to come up with a way to ensure:
shutdown pmon docker ahead of syncd for "config reload" or "service swss restart" flow;
don't shutdown pmon docker ahead of syncd for "fast-reboot" or "warm-reboot" flow in order to save time.
for "reboot" flow, either order is acceptable.
Solution
To solve the issue, pmon shoud be stopped ahead of syncd stopped for all flows except for the warm-reboot.
- How I did it
To stop pmon ahead of syncd stopped. This is done in /usr/local/bin/syncd.sh::stop() and for all shutdown sequence.
Now pmon stops ahead of syncd so there must be a way in which pmon can start after syncd started. Another point that should be taken consideration is that pmon starting should be deferred so that services which have the logic of graceful restart in fast-reboot and warm-reboot have sufficient CPU cycles to meet their deadline.
This is done by add "syncd.service" as "After" to pmon.service and startin /usr/local/bin/syncd.sh::wait()
To start pmon automatically after syncd started.
2019-09-27 03:15:46 -05:00
|
|
|
if [[ x$sonic_asic_platform == x"mellanox" ]] && [[ x$TYPE == x"cold" ]]; then
|
|
|
|
debug "Stopping pmon service ahead of syncd..."
|
|
|
|
/bin/systemctl stop pmon
|
|
|
|
debug "Stopped pmon service"
|
|
|
|
fi
|
|
|
|
|
2018-12-15 13:36:12 -06:00
|
|
|
if [[ x$sonic_asic_platform != x"mellanox" ]] || [[ x$TYPE != x"cold" ]]; then
|
|
|
|
debug "${TYPE} shutdown syncd process ..."
|
|
|
|
/usr/bin/docker exec -i syncd /usr/bin/syncd_request_shutdown --${TYPE}
|
2018-10-01 21:01:04 -05:00
|
|
|
|
2018-12-15 13:36:12 -06:00
|
|
|
# wait until syncd quits gracefully
|
|
|
|
while docker top syncd | grep -q /usr/bin/syncd; do
|
|
|
|
sleep 0.1
|
|
|
|
done
|
2018-11-15 17:47:33 -06:00
|
|
|
|
2018-12-15 13:36:12 -06:00
|
|
|
/usr/bin/docker exec -i syncd /bin/sync
|
|
|
|
debug "Finished ${TYPE} shutdown syncd process ..."
|
|
|
|
fi
|
2018-10-01 21:01:04 -05:00
|
|
|
|
2018-09-24 18:35:01 -05:00
|
|
|
/usr/bin/${SERVICE}.sh stop
|
|
|
|
debug "Stopped ${SERVICE} service..."
|
|
|
|
|
2019-01-10 16:09:04 -06:00
|
|
|
# platform specific tasks
|
|
|
|
|
2018-09-24 18:35:01 -05:00
|
|
|
if [[ x"$WARM_BOOT" != x"true" ]]; then
|
2019-02-15 18:08:08 -06:00
|
|
|
if [ x$sonic_asic_platform == x'mellanox' ]; then
|
|
|
|
/etc/init.d/sxdkernel stop
|
|
|
|
/usr/bin/mst stop
|
|
|
|
elif [ x$sonic_asic_platform == x'cavium' ]; then
|
2018-09-24 18:35:01 -05:00
|
|
|
/etc/init.d/xpnet.sh stop
|
|
|
|
/etc/init.d/xpnet.sh start
|
|
|
|
fi
|
|
|
|
fi
|
|
|
|
|
|
|
|
unlock_service_state_change
|
|
|
|
}
|
|
|
|
|
|
|
|
case "$1" in
|
2019-03-08 12:59:41 -06:00
|
|
|
start|wait|stop)
|
2018-09-24 18:35:01 -05:00
|
|
|
$1
|
|
|
|
;;
|
|
|
|
*)
|
2019-03-08 12:59:41 -06:00
|
|
|
echo "Usage: $0 {start|wait|stop}"
|
2018-09-24 18:35:01 -05:00
|
|
|
exit 1
|
|
|
|
;;
|
|
|
|
esac
|