Loggregator not accessible, errors in the logs


Rohit Kumar
 

Answers to your questions:

1. That error might be caused if your doppler servers aren't healthy and
therefore not listed in etcd. The traffic-controller polls etcd to discover
dopplers.
2. You do need metron and dopplers to access logs via "cf logs". The logs
are emitted by the DEA agent to metron, which forwards them to doppler
servers. Doppler servers deal with buffering logs (so that they are
accessible via the "cf logs --recent" command) and also deal with syslog
forwarding.

Thanks,
Rohit

On Wed, Sep 9, 2015 at 2:33 PM, kyle havlovitz <kylehav(a)gmail.com> wrote:

I'm seeing this message repeating over and over in the traffic controller
logs:
{"timestamp":1441830145.719944000,"process_id":32413,"source":"loggregator
trafficcontroller","log_level":"debug","message":"ServerAddressList.Run:
Unable to recursively find keys with prefix
/healthstatus/doppler","data":null,"file":"/opt/cloudfoundry/cf-release/src/loggregator/src/
github.com/cloudfoundry/loggregatorlib/servicediscovery/servicediscovery.go
","line":78,"method":"
github.com/cloudfoundry/loggregatorlib/servicediscovery.(*serverAddressList).DiscoverAddresses
"}

Here's my config:

"EtcdUrls" : ["http://localhost:4001"],
"EtcdMaxConcurrentRequests" : 10,
"WSMessageBufferSize": 100,
"OutgoingDropsondePort": 8082,
"DopplerPort": 8082,
"IncomingPort": 3456,
"OutgoingPort": 9090,
"SkipCertVerify": true,
"Index": 0,
"MaxRetainedLogMessages": 100,
"SharedSecret": "secret",
"Host": "0.0.0.0",
"SystemDomain": "local.example.com",
"ApiHost": "http://127.0.0.1:8181",

"NatsHosts": ["127.0.0.1"],
"NatsPort": 4222,
"NatsUser": "nats",
"NatsPass": "password",
"VarzUser": "varz",
"VarzPass": "password",
"VarzPort": 8888,
"MetronPort": 3457,
"Syslog" : "",
"Zone": "CRAZY_TOWN",

"UaaHost": "http://localhost:8080",
"UaaClientId": "doppler",
"UaaClientSecret": "tokensecret"
}

I have all the components running locally, using cf-release 215. I have
two questions:
1. what's causing this error?
2. if all I want out of the system are logs (via cf logs command), can I
get away with just having a traffic controller and dea agent running or do
I need metron and doppler?


kyle havlovitz <kylehav@...>
 

I'm seeing this message repeating over and over in the traffic controller
logs:
{"timestamp":1441830145.719944000,"process_id":32413,"source":"loggregator
trafficcontroller","log_level":"debug","message":"ServerAddressList.Run:
Unable to recursively find keys with prefix
/healthstatus/doppler","data":null,"file":"/opt/cloudfoundry/cf-release/src/loggregator/src/
github.com/cloudfoundry/loggregatorlib/servicediscovery/servicediscovery.go
","line":78,"method":"
github.com/cloudfoundry/loggregatorlib/servicediscovery.(*serverAddressList).DiscoverAddresses
"}

Here's my config:

"EtcdUrls" : ["http://localhost:4001"],
"EtcdMaxConcurrentRequests" : 10,
"WSMessageBufferSize": 100,
"OutgoingDropsondePort": 8082,
"DopplerPort": 8082,
"IncomingPort": 3456,
"OutgoingPort": 9090,
"SkipCertVerify": true,
"Index": 0,
"MaxRetainedLogMessages": 100,
"SharedSecret": "secret",
"Host": "0.0.0.0",
"SystemDomain": "local.example.com",
"ApiHost": "http://127.0.0.1:8181",

"NatsHosts": ["127.0.0.1"],
"NatsPort": 4222,
"NatsUser": "nats",
"NatsPass": "password",
"VarzUser": "varz",
"VarzPass": "password",
"VarzPort": 8888,
"MetronPort": 3457,
"Syslog" : "",
"Zone": "CRAZY_TOWN",

"UaaHost": "http://localhost:8080",
"UaaClientId": "doppler",
"UaaClientSecret": "tokensecret"
}

I have all the components running locally, using cf-release 215. I have two
questions:
1. what's causing this error?
2. if all I want out of the system are logs (via cf logs command), can I
get away with just having a traffic controller and dea agent running or do
I need metron and doppler?