We have been running fleet for a couple of months, starting with a stack running 7.15.0. We have upgraded our stack (including Kibana) to 7.16.2, and now want to upgrade our elastic-agents.
On the fleet console, our agents are showing as "Upgrade available" but the fleet server is not. If we try to upgrade a remote agent, it says its going to upgrade to 7.16.2, but the upgrade agent never reconnects to the fleet server.
If I try and upgrade the fleet server from the UI, it says its going to upgrade to 7.16.2, but then fails, with the really helpful error message that says upgrade failed.
Is there some special sauce, or magic incantation required before I can upgrade a fleet server, and then the remote agents?
Sorry to hear you're having trouble. Just want to note that the remote upgrade feature is still in beta so we still have some improvements to make (like better error messages), but in general it should work.
You will need to upgrade Fleet Server before upgrading the other Agents. One thing I'd suggest is turn on verbose Agent logging on the Agent that is running Fleet Server, and then attempting the upgrade again. Then you should be able to see more detailed information about why the upgrade may be failing in the Agent logs for your Fleet Server agent. You can do this by changing the log level drop down in the bottom right corner of the Agent detail view:
Thanks for posting the error message. Depending on how the Agent is installed, it may not eligible for remote upgrades. See the docs for more information, but in general, remote upgrades are not supported for Agents that are installed via DEB or RPM packages, or Docker containers since those are managed externally.
If you'd like to be able remote upgrade an Agent, we recommend installing via our binary tar.gz distributable directly.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.
