1.13 upgrade possible breaking changes?

Over the weekend we had an issue where a single index hit the 1.28TB limit and data stopped flowing into the cluster. Upon further inspection I noticed that the index had not been rolled over or was included in the list of managed indexes.
After creating a new index with the proper aliases and settings I noticed that data was still not flowing into the index.
Re-reading through the docs I found:

PUT _opendistro/_ism/policies/policy_id
{
  "policy": {
    "description": "Example policy.",
    "default_state": "...",
    "states": [...],
    "ism_template": {
      "index_patterns": ["index_name-*"],
      "priority": 100
    }
  }
}

Specifically that the older policy template did not include the “ism_template” block. I checked the last few change logs but did not see any mention of that of becoming a requirement.
Has anyone seen anything similar?

Once I modified my policy to include that “ism_template” block everything started working as expected.

Hey @jasonrojas,

It’s included in the list of breaking changes in the release notes:

It did get buried under a big table though… but, as you found out we removed support of the policy_id index setting and introduced the usage of ISM templates instead.

2 Likes

Ahh. Sure enough, there it is. Thanks!