- NEW: Upgrade fails on installations with external PostgreSQL
- NEW: PostgreSQL upgrade fails if a package provider is broken
- Ubuntu 16.10 agents can't be installed with package management
- Installer can fail due to SSL errors with AmazonAWS
- Incorrect credentials for console databases cause split upgrade to fail
- Web-based installer fails to acknowledge failed installs due to low RAM
- Hard tabs for indentation in Hiera YAML files cause errors after upgrading
- Incorrect umask value can cause installation and upgrade to fail
- Upgrade fails if autosign.conf contains invalid entries
- Install agents with different OS when master is behind a proxy
These are the known issues for installation and upgrade in this release.
NEW: Upgrade fails on installations with external PostgreSQL
When upgrading a split installation with an external PostgreSQL node, the installation fails and reverts all package upgrades because the PostgreSQL version can't be validated.
NEW: PostgreSQL upgrade fails if a package provider is broken
PostgreSQL is upgraded to version 9.6, but data fails to migrate if the
pe_postgresql_info custom fact in the
pe_install module fails to resolve. This issue is usually caused by a broken Puppet package provider.
Ubuntu 16.10 agents can't be installed with package management
Installation using package management isn't currently supported. If you need to install an Ubuntu 16.10 agent, log into the node where you want to install the agent and run:
wget https://apt.puppetlabs.com/puppetlabs-release-pc1-yakkety.deb sudo dpkg -i puppetlabs-release-pc1-yakkety.deb
Installer can fail due to SSL errors with AmazonAWS
In some cases when attempting to install, some master platforms have received SSL errors attempting to connect to s3.amazonaws.com. Consequently, agent packages needed for installation can't be retrieved. In most cases, you can successfully install after updating the CA cert bundle on the master platform. To update the bundle, run:
rm /etc/ssl/certs/ca-bundle.crt yum reinstall ca-certificates
After updating the CA bundle, run the installer again.
Incorrect credentials for console databases cause split upgrade to fail
During a split upgrade, if you supply incorrect credentials for the database associated with the console, including database names, user names, or passwords, the upgrade process fails with an error message. Verify that you're using the correct database credentials and re-run the upgrader. The credentials can be found on the PuppetDB node at
Web-based installer fails to acknowledge failed installs due to low RAM
When installation fails because a system is not provisioned with adequate RAM, the web-based installer stops responding, but the Start using Puppet Enterprise button is available, suggesting that installation succeeded. In these cases, the command line shows an "out of memory: Kill process" error.
Provision the system with adequate RAM and re-run the installation.
Hard tabs for indentation in Hiera YAML files cause errors after upgrading
Before upgrade, ensure that any Hiera YAML files do not use hard tabs for indentation.
Incorrect umask value can cause installation and upgrade to fail
Set an umask value of
0022 on your master.
Upgrade fails if autosign.conf contains invalid entries
/etc/puppetlabs/puppet/autosign.conf that don't conform to autosign requirements cause the upgrade to fail when configuring the console. Correct any invalid entries before upgrading.
Install agents with different OS when master is behind a proxy
If your master uses a proxy server to access the internet, you might not be able to download
pe_repo packages for the agent.
If you're using a proxy, follow this workaround:
- From your master, navigate to
/etc/sysconfig/, and create a file called
puppetwith this code:
export http_proxy <YOUR_PROXY_SERVER> export https_proxy <YOUR_PROXY_SERVER>
Save and close the file, and then restart the
puppet resource service puppet ensure=stopped puppet resource service puppet ensure=running
Repeat these steps on any compile masters in your environment.