[7.1.1] Upgrade docs error
abraxxa opened this issue · 6 comments
Role version: 7.1.1
Description of the problem including expected versus actual behaviour:
The upgrade docs lists the four variables that need to be set to the existing directories when upgrading from a previous playbook version to 7.1.1.
The variable instance_name was called es_instance_name, see tag 6.6.0 for example.
As node.name defaults to the hostname using the ansible variable ansible_hostname might be a better recommendation instead of node_name which is undefined when using the previous playbook version.
Hi @abraxxa,
Thanks a lot for the issue, {{ node_name }} and {{ instance_name }} in the upgrade doc were more placeholders to replace than real ansible variables but you're totally right, it's more consistent and clearer to use ansible variables.
For node_name, I would use inventory_hostname instead of ansible_hostname to be fully consistent with previous behavior (https://github.com/elastic/ansible-elasticsearch/pull/566/files#diff-916074b85953579021fe6cee52c54d8bL65).
Would you mind creating a PR for that change?
Yeah I thought so but providing usable ones if available is even better
I have submitted PRs in the past which couldn't be merged because I can't sign the CLA, so please create it yourself.
Thanks!
You're right on instance_name as the previous playbook version used the fqdn for the node name, not just its hostname, so inventory_hostname works for me but might not if e.g. only the hostname is used for the inventory and ansible_host is defined.
And another issue I've encountered which fits this bug report as well:
LimitMEMLOCK=infinity isn't defined in the systemd service file of the Debian package and the playbook doesn't generate a /etc/systemd/system/elasticsearch.service.d/override.conf containing
[Service]
LimitMEMLOCK=infinity
as the docs suggest.
Should be generated by the playbook as well.