istio: envoy crash when setup envoy access log service

Bug description

after change istio cm as bellow:

envoy crash and get logs as follow:

2021-08-27T06:34:45.956129Z	critical	envoy main	std::terminate called! (possible uncaught exception, see trace)
2021-08-27T06:34:45.956159Z	critical	envoy backtrace	Backtrace (use tools/stack_decode.py to get line numbers):
2021-08-27T06:34:45.956160Z	critical	envoy main	std::terminate called! (possible uncaught exception, see trace)
2021-08-27T06:34:45.956179Z	critical	envoy backtrace	Backtrace (use tools/stack_decode.py to get line numbers):
2021-08-27T06:34:45.956182Z	critical	envoy backtrace	Envoy version: 9a976952980250ed356ebdd4f94e219e28a041dd/1.19.1-dev/Clean/RELEASE/BoringSSL
2021-08-27T06:34:45.956163Z	critical	envoy backtrace	Envoy version: 9a976952980250ed356ebdd4f94e219e28a041dd/1.19.1-dev/Clean/RELEASE/BoringSSL
2021-08-27T06:34:45.961294Z	warning	envoy config	gRPC config for type.googleapis.com/envoy.config.listener.v3.Listener rejected: Error adding/updating listener(s) 10.237.207.202_15012: Unknown gRPC client cluster 'envoy_accesslog_service'
10.237.0.1_443: Unknown gRPC client cluster 'envoy_accesslog_service'
10.237.0.10_53: Unknown gRPC client cluster 'envoy_accesslog_service'
10.237.207.202_443: Unknown gRPC client cluster 'envoy_accesslog_service'
0.0.0.0_80: Unknown gRPC client cluster 'envoy_accesslog_service'
0.0.0.0_8080: Unknown gRPC client cluster 'envoy_accesslog_service'
10.237.0.209_8000: Unknown gRPC client cluster 'envoy_accesslog_service'
0.0.0.0_15010: Unknown gRPC client cluster 'envoy_accesslog_service'
10.237.0.10_8080: Unknown gRPC client cluster 'envoy_accesslog_service'
0.0.0.0_15014: Unknown gRPC client cluster 'envoy_accesslog_service'
virtualOutbound: Unknown gRPC client cluster 'envoy_accesslog_service'
virtualInbound: Unknown gRPC client cluster 'envoy_accesslog_service'

2021-08-27T06:34:45.979164Z	critical	envoy backtrace	#0: Envoy::TerminateHandler::logOnTerminate()::$_0::operator()() [0x5617e6f01486]
2021-08-27T06:34:45.979177Z	critical	envoy backtrace	#0: Envoy::TerminateHandler::logOnTerminate()::$_0::operator()() [0x5617e6f01486]
2021-08-27T06:34:45.987414Z	critical	envoy backtrace	#1: [0x5617e6f012e9]
2021-08-27T06:34:45.987428Z	critical	envoy backtrace	#1: [0x5617e6f012e9]
2021-08-27T06:34:45.995368Z	critical	envoy backtrace	#2: std::__terminate() [0x5617e75a9633]
2021-08-27T06:34:45.995386Z	critical	envoy backtrace	#2: std::__terminate() [0x5617e75a9633]
2021-08-27T06:34:46.002841Z	critical	envoy backtrace	#3: Envoy::Grpc::AsyncClientManagerImpl::factoryForGrpcService() [0x5617e674a74a]
2021-08-27T06:34:46.002866Z	critical	envoy backtrace	#3: Envoy::Grpc::AsyncClientManagerImpl::factoryForGrpcService() [0x5617e674a74a]
2021-08-27T06:34:46.010364Z	critical	envoy backtrace	#4: Envoy::Extensions::AccessLoggers::Common::GrpcAccessLoggerCache<>::getOrCreateLogger() [0x5617e4fe181d]
2021-08-27T06:34:46.012595Z	critical	envoy backtrace	#4: Envoy::Extensions::AccessLoggers::Common::GrpcAccessLoggerCache<>::getOrCreateLogger() [0x5617e4fe181d]
2021-08-27T06:34:46.018062Z	critical	envoy backtrace	#5: Envoy::Extensions::AccessLoggers::Common::GrpcAccessLoggerCache<>::getOrCreateLogger() [0x5617e4fe1a32]
2021-08-27T06:34:46.019678Z	critical	envoy backtrace	#5: Envoy::Extensions::AccessLoggers::Common::GrpcAccessLoggerCache<>::getOrCreateLogger() [0x5617e4fe1a32]
2021-08-27T06:34:46.025081Z	critical	envoy backtrace	#6: std::__1::__function::__func<>::operator()() [0x5617e4fe2e7d]
2021-08-27T06:34:46.026669Z	critical	envoy backtrace	#6: std::__1::__function::__func<>::operator()() [0x5617e4fe2e7d]
2021-08-27T06:34:46.032533Z	critical	envoy backtrace	#7: std::__1::__function::__func<>::operator()() [0x5617e66bb918]
2021-08-27T06:34:46.033677Z	critical	envoy backtrace	#7: std::__1::__function::__func<>::operator()() [0x5617e66bb918]
2021-08-27T06:34:46.040005Z	critical	envoy backtrace	#8: Envoy::Event::DispatcherImpl::runPostCallbacks() [0x5617e6c982cd]
2021-08-27T06:34:46.040706Z	critical	envoy backtrace	#8: Envoy::Event::DispatcherImpl::runPostCallbacks() [0x5617e6c982cd]
2021-08-27T06:34:46.047389Z	critical	envoy backtrace	#9: event_process_active_single_queue [0x5617e6dbd6d8]
2021-08-27T06:34:46.047735Z	critical	envoy backtrace	#9: event_process_active_single_queue [0x5617e6dbd6d8]
2021-08-27T06:34:46.054742Z	critical	envoy backtrace	#10: event_base_loop [0x5617e6dbc3c1]
2021-08-27T06:34:46.055113Z	critical	envoy backtrace	#10: event_base_loop [0x5617e6dbc3c1]
2021-08-27T06:34:46.062208Z	critical	envoy backtrace	#11: Envoy::Server::WorkerImpl::threadRoutine() [0x5617e66d958d]
2021-08-27T06:34:46.063583Z	critical	envoy backtrace	#11: Envoy::Server::WorkerImpl::threadRoutine() [0x5617e66d958d]
2021-08-27T06:34:46.069610Z	critical	envoy backtrace	#12: Envoy::Thread::ThreadImplPosix::ThreadImplPosix()::{lambda()#1}::__invoke() [0x5617e6fea673]
2021-08-27T06:34:46.069692Z	critical	envoy backtrace	#13: start_thread [0x7ff20dd6f609]
2021-08-27T06:34:46.069719Z	critical	envoy backtrace	Caught Aborted, suspect faulting address 0x53900000011
2021-08-27T06:34:46.069725Z	critical	envoy backtrace	Backtrace (use tools/stack_decode.py to get line numbers):
2021-08-27T06:34:46.069727Z	critical	envoy backtrace	Envoy version: 9a976952980250ed356ebdd4f94e219e28a041dd/1.19.1-dev/Clean/RELEASE/BoringSSL
2021-08-27T06:34:46.069757Z	critical	envoy backtrace	#0: __restore_rt [0x7ff20dd7b3c0]
2021-08-27T06:34:46.070785Z	critical	envoy backtrace	#12: Envoy::Thread::ThreadImplPosix::ThreadImplPosix()::{lambda()#1}::__invoke() [0x5617e6fea673]
2021-08-27T06:34:46.070843Z	critical	envoy backtrace	#13: start_thread [0x7ff20dd6f609]
2021-08-27T06:34:46.074023Z	info	ads	ADS: "@" nginx-69b486cdfd-vjsm8.default-1 terminated rpc error: code = Canceled desc = context canceled
2021-08-27T06:34:46.074089Z	info	ads	ADS: "@" nginx-69b486cdfd-vjsm8.default-2 terminated rpc error: code = Canceled desc = context canceled
2021-08-27T06:34:46.074287Z	error	Epoch 0 exited with error: signal: aborted
2021-08-27T06:34:46.074309Z	info	No more active epochs, terminating

Affected product area (please put an X in all that apply)

[ ] Docs [ ] Installation [ ] Networking [ ] Performance and Scalability [ ] Extensions and Telemetry [ ] Security [ ] Test and Release [ ] User Experience [ ] Developer Infrastructure [ ] Upgrade

Affected features (please put an X in all that apply)

[ ] Multi Cluster [ ] Virtual Machine [ ] Multi Control Plane

Expected behavior

envoyAccessLogService not work before re-inject the pod, but the envoy should not crash

Steps to reproduce the bug

• install mesh istioctl install --set profile=minimal
• inject a pod in the default namespace
• change the istio cm in istio-system namespace

then you will the pod envoy crashed

istioctl install --set profile=minimal

Version (include the output of istioctl version --remote and kubectl version --short and helm version --short if you used Helm)

istioctl

client version: 1.11.1
control plane version: 1.11.1
data plane version: 1.11.1 (1 proxies)

kubectl

Client Version: v1.17.11
Server Version: v1.17.17-r0-CCE21.6.1.B004-17.37.5

How was Istio installed?

istioctl

Environment where the bug was observed (cloud vendor, OS, etc)

Additionally, please consider running istioctl bug-report and attach the generated cluster-state tarball to this issue. Refer cluster state archive for more details.