Graylog login loops back

Graylog Central
#1

I have a running graylog service, I have one master and one slave. The problem is that the login screen refreshes when I log in and put my credentials. It again asks me to login without giving any error. Both my master and slave have the same secret password and Sha2 password is correct. When I only run the master, everything runs fine, I am able to login. But not with the slave.
Graylog Version : 2.2.3-1

EDIT : Sometimes Chrome shows this error:

Error: cannot GET http://104.154.244.243:31300/api/system/inputs (401)

Error: cannot GET http://104.154.244.243:31300/api/streams (401)

Thats all, No error whatsoever in any other logs.
Please help.

#2

How did you install Graylog?
How did you configure both Graylog nodes, especially the rest_listen_uri, rest_transport_uri, and web_endpoint_uri settings?

1 Like
#3

I am directly putting the whole configuration so that no doubt is there on the variables. Please look into.
These are the part of my configuration for the master and slave

Master :

   name: graylog-master
          image: graylog2/server:2.2.3-1
          env:
          - name: GRAYLOG_PASSWORD_SECRET
            value: "somesaltpassword"
          - name: GRAYLOG_ROOT_PASSWORD_SHA2
            value: "5D5E792708BFA15F0AB42E817B4E69379777D2722E0529DFB031C0B847DB137D"
          - name: GRAYLOG_MONGODB_URI
            value: "mongodb://mongo-mongodb-replicaset:27017/graylog"
          - name: GRAYLOG_IS_MASTER
            value: "true"
          - name: GRAYLOG_ELASTICSEARCH_DISCOVERY_ZEN_PING_UNICAST_HOSTS
            value: "elasticsearch-0.elasticsearch.default.svc.cluster.local:9300, elasticsearch-1.elasticsearch.default.svc.cluster.local:9300, elasticsearch-2.elasticsearch.default.svc.cluster.local:9300"
          - name: GRAYLOG_WEB_ENDPOINT_URI
            value: "http://104.154.244.243:31300/api"
          - name: GRAYLOG_REST_LISTEN_URI
            value: "http://0.0.0.0:9000/api/"
          - name: GRAYLOG_WEB_LISTEN_URI
            value: "http://0.0.0.0:9000/"
          - name: GRAYLOG_REST_TRANSPORT_URI
            value: "http://graylog-master.default.svc.cluster.local:9000/api/"
          ports:
          - containerPort: 9000
            name: http
          - containerPort: 12201
            name: udp-input
          - containerPort: 1514
            name: tcp-input

Slave:

name: graylog-slave
          image: graylog2/server:2.2.3-1
          env:
          - name: GRAYLOG_PASSWORD_SECRET
            value: "somesaltpassword"
          - name: GRAYLOG_ROOT_PASSWORD_SHA2
            value: "5D5E792708BFA15F0AB42E817B4E69379777D2722E0529DFB031C0B847DB137D"
          - name: GRAYLOG_MONGODB_URI
            value: "mongodb://mongo-mongodb-replicaset:27017/graylog"
          - name: GRAYLOG_IS_MASTER
            value: "false"
          - name: GRAYLOG_ELASTICSEARCH_DISCOVERY_ZEN_PING_UNICAST_HOSTS
            value: "elasticsearch-0.elasticsearch.default.svc.cluster.local:9300, elasticsearch-1.elasticsearch.default.svc.cluster.local:9300, elasticsearch-2.elasticsearch.default.svc.cluster.local:9300"
          - name: GRAYLOG_WEB_ENDPOINT_URI
            value: "http://104.154.244.243:31300/api"
          - name: GRAYLOG_REST_LISTEN_URI
            value: "http://0.0.0.0:9000/api/"
          - name: GRAYLOG_WEB_LISTEN_URI
            value: "http://0.0.0.0:9000/"
          - name: GRAYLOG_REST_TRANSPORT_URI
            value: "http://graylog-slave.default.svc.cluster.local:9000/api/"
          ports:
          - containerPort: 9000
            name: http
          - containerPort: 12201
            name: udp-input
          - containerPort: 1514
            name: tcp-input
```

Note : graylog-slave and graylog-master are the headless services. The mongodb and elasticsearch have their own stateful sets and I am using gce for this work. Albeit, no error in the logs against connection of them.

The master and slave headless are almost the same and look as : 
```
spec:
  selector:
    role: graylog-master (slave in the other case)
  ports:
  - name: http
    protocol: TCP
    port: 9000
  clusterIP: None
```

Please reply.
#4

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.