Error: "Version does not adhere to semantic versioning" when running DataDog agent 6.16.0 on Ubuntu
See original GitHub issueDescribe what happened:
After upgrading the DataDog agent to version 6.16.0
, we are getting an agent error like the following posted to stdout about 10 times every second. This is causing a lot of noise on our logs.
2019-12-19 20:20:23 UTC | CORE | ERROR | (pkg/collector/python/datadog_agent.go:116 in LogMessage) | (core.py:45) | Unable to transform `version` metadata value `Shield`: Version does not adhere to semantic versioning
We follow the default installation steps, just by executing curl -L https://raw.githubusercontent.com/DataDog/datadog-agent/master/cmd/agent/install_script.sh
on Ubuntu
The following can be seen in the installation logs - perhaps the version string 6.16.0-1
is the cause of the SemVer error?
remote: Preparing to unpack .../datadog-agent_1%3a6.16.0-1_amd64.deb ...
remote: INFO -- : Unpacking datadog-agent (1:6.16.0-1) ...
remote: ....
remote: INFO -- : Processing triggers for systemd (229-4ubuntu21.23) ...
remote: INFO -- : Setting up datadog-agent (1:6.16.0-1) ...
Describe what you expected:
No version errors posted on stdout/stderr
Steps to reproduce the issue:
Install Datadog agent using curl -L https://raw.githubusercontent.com/DataDog/datadog-agent/master/cmd/agent/install_script.sh
and start the agent using /usr/bin/datadog-agent run
.
Additional environment details (Operating System, Cloud provider, etc):
Ubuntu Bionic running on Docker and hosted on AWS
Issue Analytics
- State:
- Created 4 years ago
- Reactions:1
- Comments:5 (3 by maintainers)
Top GitHub Comments
Hey @therve that’s it! We were blocking
server_tokens
in the main config so added it to the private status location and that did the trick! Errors and noise are gone from the logs - thanks for your help! Feel free to close the issueThanks for response! It looks like the issue we fixed at https://github.com/DataDog/integrations-core/pull/5250 albeit with a different error. One thing you can try is to set
server_tokens on
on the status location, if that endpoint is private. That would allow us to collect the version properly. We also have another patch to reduce the level of that log.