ERR: (/Stage[main]/Mongodb::Replset/Mongodb_replset[ceilometer]) Could not evaluate: rs.initiate() failed for replicaset ceilometer: not authorized on admin to execute command
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Released
|
High
|
Ivan Berezovskiy | ||
8.0.x |
Won't Fix
|
High
|
Ivan Berezovskiy |
Bug Description
1. Create new environment
2. Choose Neutron, tunnelling segmentation
3. Choose Ceilometer
4. Add 3 controller+mongo, 1 compute, 1 cinder, 3 mongo
5. Change disk configuration for all Mongo nodes. Change 'MongoDB' volume for vdc
6. Start deployment. Controller and mongo nodes in 'Error' state. Deployment has failed.
Error in puppet.log:
2016-01-27 10:32:37 ERR (/Stage[
This error appeared on primary controller (node-8), mongo nodes (node-4,5)
VERSION:
feature_groups:
- mirantis
production: "docker"
release: "8.0"
api: "1.0"
build_number: "478"
build_id: "478"
fuel-nailgun_sha: "ae949905142507
python-
fuel-agent_sha: "481ed135de2cb5
fuel-
astute_sha: "b81577a5b7857c
fuel-library_sha: "420c6fa5f8cb51
fuel-ostf_sha: "ab5fd151fc6c1a
fuel-mirror_sha: "b62f3cce5321fd
fuelmenu_sha: "fac143f4dfa757
shotgun_sha: "63645dea384a37
network-
fuel-upgrade_sha: "616a7490ec7199
fuelmain_sha: "6c6b088a3d52dd
description: | updated |
tags: | added: area-puppet |
tags: | added: swarm-blocker |
tags: | added: on-verification |
I've researched environment and found root cause of this problem. Admin user wasn't created, because replica initialization took a lot of time. 10 retires wasn't enough to bring replica set up. After mongodb logs analysis I can say that the problem in very slow environment (weak vm for mongodb node). This retries should be improved later: timeout should not be just increased, we should be able to configure it somehow and pass into mongodb_replset provider.