Strimzi
Replies: 2 comments 5 replies
-
|
I think you will need to share the full logs from the relevant components. Especially the Topic Operator, its TLS sidecar, and ZooKeeper. Are you using some special IPv6 networking setup? The log seems to be referring to |
Beta Was this translation helpful? Give feedback.
-
|
thank you Jakub, it seems not finish this sync step, could you let me know how can I change this default vault based on my setting below, i would like have a try , |
Beta Was this translation helpful? Give feedback.
-
|
I'm facing similar issue. For some reason, entityOperator container is running with: STRIMZI_ZOOKEEPER_CONNECT=localhost:2181 spec:
containers:
- args:
- /opt/strimzi/bin/topic_operator_run.sh
env:
- name: STRIMZI_RESOURCE_LABELS
value: strimzi.io/cluster=kafka-test
- name: STRIMZI_KAFKA_BOOTSTRAP_SERVERS
value: 'kafka-test-kafka-bootstrap:9091'
- name: STRIMZI_NAMESPACE
value: kafka
- name: STRIMZI_ZOOKEEPER_CONNECT
value: 'localhost:2181'
- name: STRIMZI_ZOOKEEPER_SESSION_TIMEOUT_MS
value: '18000'
- name: STRIMZI_TOPIC_METADATA_MAX_ATTEMPTS
value: '6'
- name: STRIMZI_FULL_RECONCILIATION_INTERVAL_MS
value: '60000'
- name: STRIMZI_SECURITY_PROTOCOL
value: SSL
- name: STRIMZI_TLS_ENABLED
value: 'true'
- name: STRIMZI_GC_LOG_ENABLED
value: 'false'Failing with: 2024-09-26 10:05:32,56158 ERROR [vert.x-eventloop-thread-0] Session:246 - Topic operator start up failed, cause was
org.I0Itec.zkclient.exception.ZkTimeoutException: Unable to connect to zookeeper server 'localhost:2181' with timeout of 18000 msStrange thing is, tls-sidecar running in the same pod at entityOperator has a different value in STRIMZI_ZOOKEEPER_CONNECT variable. - args:
- /opt/stunnel/entity_operator_stunnel_run.sh
env:
- name: TLS_SIDECAR_LOG_LEVEL
value: notice
- name: STRIMZI_ZOOKEEPER_CONNECT
value: 'kafka-test-kafka-zookeeper-client:2181'
image: 'strimzi/kafka:0.37.0-kafka-3.5.1'
name: tls-sidecarI only used these additional parameters to configure entityOperator. Including my attempts to override STRIMZI_ZOOKEEPER_CONNECT variable, which does not work. spec:
clusterCa:
generateCertificateAuthority: true
entityOperator:
template:
pod:
securityContext:
fsGroup: 3000
runAsGroup: 2000
runAsUser: 10001
topicOperatorContainer:
env:
- name: STRIMZI_ZOOKEEPER_CONNECT
value: 'kafka-test-kafka-zookeeper-client:2181'I'm not sure how these environment variables are populated by the strimzi cluster operator. |
Beta Was this translation helpful? Give feedback.
-
|
You should not touch these environment variables.
In this case, the Topic Operator connects through localhost to the TLS sidecar that proxies the connection to the actual Zookeeper and handles the security. So for the old Topic Operator, connecting to |
Beta Was this translation helpful? Give feedback.
-
|
Alright, makes perfect sense so the actual error lands into the tls sidecar container. kind: Kafka
spec:
clusterCa:
generateCertificateAuthority: trueAll component got restarted fine but I'm not sure yet why this is failing. |
Beta Was this translation helpful? Give feedback.
-
|
Have you tried to restart it manually? |
Beta Was this translation helpful? Give feedback.
-
|
yes but it does not seem to work. Enabling clusterCA generation and restarting all components caused a failure because Thanks for your support Jakub |
Beta Was this translation helpful? Give feedback.
-
Hello,
I have deployed a version 0.35.1 cluster operator with 1 kafka broker and 1 zookeeper
from pod status, kafka broker and zookeeper running normally, and checked kafka and zookeeper log both are normal,
but entity-operator still 2/3 which Topic operator container start failed ,
ERROR [vert.x-eventloop-thread-1] Main:68 - Error deploying Session
org.I0Itec.zkclient.exception.ZkTimeoutException: Unable to connect to zookeeper server 'localhost:2181' with timeout of 18000 ms
below are the information , please let me know if need more,
btw , tried producer consumer test, looks working
thanks in advance
Beta Was this translation helpful? Give feedback.
All reactions