Marvel license expired, new license is already expired

monitoring

(Steven) #1

I noticed that my marvel license was expired. So I registered for a new Basic license, got an email, downloaded it, and ran the command

curl -XPUT 'http://es01:9200/_license' -d @license.json

Looks good http://es01:9200/_license:

{
  "license": {
    "status": "active",
    "uid": "<long-id-here>",
    "type": "basic",
    "issue_date": "2016-06-27T00:00:00.000Z",
    "issue_date_in_millis": 1466985600000,
    "expiry_date": "2017-07-06T23:59:59.999Z",
    "expiry_date_in_millis": 1499385599999,
    "max_nodes": 100,
    "issued_to": "<my-name-here> (<my-company-here>)",
    "issuer": "Web Form"
  }
}

But visiting http://es01:5601/app/marvel shows Basic as grayed out and Trial is expired:

Why is that?


(Chris Earle) #2

Hi Styfle,

The answer lies in how clusters are identified by Marvel (and, internally to Elasticsearch) versus how they are displayed to humans.

When you create a cluster, the first thing that it does it to create a Cluster UUID that gets written by the elected master node and shared with all other nodes in the cluster. Internally, Elasticsearch uses the Cluster UUID to guarantee cluster uniqueness rather than the arbitrary name that we deal with as people.

It's that same UUID that Marvel uses to determine cluster uniqueness for this list. We display the cluster name, because that's what people are able to actually use, but the cluster uuid is all that we care about.

What I expect has happened is that the cluster with the expired trial license was a test cluster whose data directory got wiped out, then replaced by a fresh cluster, which got its own UUID. Marvel kept track of the old one, so it prevents multi-cluster access with a Basic license (hence the grayed out cluster).

The next question is naturally:

How do I fix this?

Since your fixed cluster only has day from today, based on the license anyway, the easiest thing to do is to delete the Marvel data.

$ curl -XDELETE es01:9200/.marvel*

After less than a minute (about 10 seconds by default), the real cluster will send some fresh data and it will recreate the necessary indices.

Hope that helps,
Chris


(Steven) #3

Hi Chris,

Thank you for the well-written explanation!
Deleting the marvel data worked!

:grin:


(system) #4