You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{jsonble}
They are depreciating how _uid operates in this version, and it starting to use a _id search instead which only allows for simple queries like term searches. This will also affect our elasticsearch_reader if we use the capability to further split a time segment further down.
Also if this were applied as a subslicing technique for further division of an indivisible date range query, the impact of the state held server side would be relatively small and not held for the entire duration of a reader operation. It would only have to be held for the duration of the subslicing.
In cases where an index is static using the scroll API should be perfectly fine so it may still give us a reasonable set of options for when range slicing doesn't work.
The text was updated successfully, but these errors were encountered:
{jsonble}
They are depreciating how _uid operates in this version, and it starting to use a _id search instead which only allows for simple queries like term searches. This will also affect our elasticsearch_reader if we use the capability to further split a time segment further down.
docs:
https://www.elastic.co/guide/en/elasticsearch/reference/6.2/mapping-uid-field.html
https://www.elastic.co/guide/en/elasticsearch/reference/6.2/mapping-id-field.html
for that last one pay attention to the note posted towards the top of the page and the paragraph just past it.
{kstaken}
This appears to be the root of the problem: elastic/elasticsearch#18154
And where they decided to not worry about what this breaks: elastic/elasticsearch#24247
Just so that we maintain an option for working in this regard we may need to look at sliced scroll queries and evaluate just how much risk there is from the state held server side on active indices. https://www.elastic.co/guide/en/elasticsearch/reference/current/search-request-scroll.html
Also if this were applied as a subslicing technique for further division of an indivisible date range query, the impact of the state held server side would be relatively small and not held for the entire duration of a reader operation. It would only have to be held for the duration of the subslicing.
In cases where an index is static using the scroll API should be perfectly fine so it may still give us a reasonable set of options for when range slicing doesn't work.
The text was updated successfully, but these errors were encountered: