Total TTL confusion

I'm getting regular TTL exceptions in my elasticsearch setup during
updating documents. I want to figure out how exactly TLL works.
The more I read about it, the more confusing it gets.

I have several questions, hope anyone is willing to answer them:

*** Using elasticsearch 0.90.5
*** In my mapping I set the -tll to 30 days using "_ttl": { "enabled":
true, "default": "30d" }. I can confirm this works by querying the _mapping
endpoint. ***

  1. when I update an existing document, will the ttl be updated or not?
    and also make
    reference to the fact that the TTL will not be updated when you do an
    update request. Which makes sense since you basically do a delete and
    reindex for that document. However, on
    it states :

Note that the expiration procedure handle versioning properly so if a
document is updated between the collection of documents to expire and the
delete order, the document won’t be deleted.

Do the two resources conflict each other?

  1. Since I have a default ttl of 30days, I would expect my documents to
    have a ttl of 2592000000ms at most. However, when I query for all the TTL's
    in my index (using a statistical facet for example), the minum value I get
    is 1407373291910ms, which is over 44 years!
    states that

If no default is set and no _ttl value is given then the document has an
infinite _ttl and will not expire.

Is this what happened here? Does the update query not (correctly / at all)
set the ttl?

The expiration date that will be set for a document with a provided ttl is
relative to the timestamp of the document

However; I have NOT set a timestamp OR enabled it in my mapping. As far as
I know my documents do not have any timestamp.
How does Elasticsearch compute the expiration date on these documents?

Any help would be much appreciated here . Thank you!

You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
To view this discussion on the web visit
For more options, visit