JaredTan95
commented
1 year ago @felixgao Have you tried use non-host network in your docker-compose yaml?
Closed felixgao closed 11 months ago
JaredTan95
commented
1 year ago @felixgao Have you tried use non-host network in your docker-compose yaml?
felixgao
commented
1 year ago I have not, since I am not building a docker image for my application that uses OTEL yet, I thought it would be good to setup the OTEL infrastructure locally first with docker first. That is why I want to use the host network.
codeboten
commented
1 year ago you can try binding the otlp ports to a different port. i believe jaeger is now also binding to those ports, which would be causing the problem you're seeing.
felixgao
commented
1 year ago you can try binding the otlp ports to a different port. i believe jaeger is now also binding to those ports, which would be causing the problem you're seeing.
I think the problem is not my host OTLP port, it is the the port inside the docker container.
I have changed my host port to
2023-10-18T17:41:26.097Z info service@v0.87.0/service.go:192 Shutdown complete.
Error: cannot start pipelines: listen tcp 0.0.0.0:4317: bind: address already in use
2023/10/18 17:41:26 collector server run finished with error: cannot start pipelines: listen tcp 0.0.0.0:4317: bind: address already in useI think my compose is very much like the demo compose file with the major difference on using a host network. I could try to not use the host network.
Update: I have removed the network_mode: host and service is up and running without any errors.
codeboten
commented
1 year ago Right the configuration I was talking about is the collector configuration. The change would be to go from
receivers:
otlp:
protocols:
grpc:
http:
zipkin:to something like this, i picked arbitrary ports here
receivers:
otlp:
protocols:
grpc:
endpoint: localhost:9317
http:
endpoint: localhost:9318
zipkin:The docker compose port configuration doesn't change what port the service is binding to.
codeboten
commented
11 months ago Marking this as closed based on the comment
Update: I have removed the network_mode: host and service is up and running without any errors.
Please re-open if that's not the case.
StMakhov
commented
1 month ago Unfortunately, I have the same problem and removing network_mode: host from docker-compose.yaml didn't solve the problem for me. Could you please clarify what else might help?
I tried to receive data in OpenTelemetry collector and export it to jaeger-all-in-one. As a source of telemetry I used telemetrygen from official documentation example . I tried to start components from docker compose and using binaries, the error is the same (commands below are listed only for the second case ).
Config.yaml for the collector:
user@host:~$ cat /etc/otelcol/config.yaml
extensions:
health_check: {}
pprof: {}
zpages: {}
receivers:
otlp:
protocols:
grpc:
endpoint: localhost:4317
http:
endpoint: localhost:4318
zipkin: {}
processors:
attributes:
actions:
- key: some.ip
value: "some_IP_in_future"
action: insert
batch: {}
exporters:
debug:
verbosity: detailed
otlp:
endpoint: jaeger-all-in-one:4317
tls:
insecure: true
zipkin:
endpoint: "http://zipkin:9411/api/v2/spans"
service:
pipelines:
traces:
receivers: [otlp, zipkin]
processors: [attributes, batch]
exporters: [debug, otlp]
metrics:
receivers: [otlp]
exporters: [debug, otlp]
logs:
receivers: [otlp]
processors: [batch]
exporters: [debug]
extensions: [pprof, zpages, health_check]
telemetry:
logs:
level: "debug"Starting the collector:
user@host:~$ /usr/bin/otelcol --config=file:/etc/otelcol/config.yaml --feature-gates=-component.UseLocalHostAsDefaultHostStarting the load:
$GOBIN/telemetrygen traces --otlp-insecure --traces 3I see log messages in the terminal with OpenTelemetry collector from telemetrygen, however, jaeger-all-in-one fails with an error when I try to start it:
user@host:~/pkg/jaeger-1.61.0-linux-amd64$ ./jaeger-all-in-one --collector.otlp.enabled=true
...
could not start OTLP receiver: could not start the OTLP receiver: listen tcp :4317: bind: address already in useOpenTelemetry collector also sends an error to its terminal:
Exporting failed. Will retry the request after interval. {"kind": "exporter", "data_type": "traces", "name": "otlp", "error": "rpc error: code = Unavailable desc = name resolver error: produced zero addresses", "interval": "13.56862226s"}If I don't add any endpoints both to receivers::otlp::protocols::grpc and receivers::otlp::protocols::http, my OpenTelemetry collector instance doesn't receive data from the load, it is sent directly to Jaeger Collector and further to Jaeger UI from it, so the OpenTelemetry collector is just ignored in this case.
Component(s)
No response
What happened?
Description
I downloaded the latest docker image and trying to bring up a collect to sent my trace/metrics over. When I tried to
compose upmy services the collector is failing withSteps to Reproduce
docker-compose up -d collectorExpected Result
services up and running
Actual Result
container exited with error.
Collector version
208830cb18d9
Environment information
Environment
OS: (e.g., "Ubuntu 20.04") macOS Ventura 13.5.2 Compiler(if manually compiled): (e.g., "go 14.2") Docker version: 4.23.0
OpenTelemetry Collector configuration
No response
Log output
lsof of listening ports
Additional context
docker-compose.yaml
otel-collector-config.yaml