# \[5.6.5\] terminate\_after and the Scroll API **URL:** \ **Category:** Elasticsearch\ **Created:** [March 25, 2020, 11:43pm UTC](https://discuss.elastic.co/t/5-6-5-terminate-after-and-the-scroll-api/225083 "2020-03-25T23:43:47Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![darin](https://avatars.discourse-cdn.com/v4/letter/d/848f3c/32.png) [@darin](https://discuss.elastic.co/u/darin)\ **Post date:** [March 25, 2020, 11:43pm UTC](https://discuss.elastic.co/t/5-6-5-terminate-after-and-the-scroll-api/225083/1 "2020-03-25T23:43:47Z")
All, I'm seeing some peculiar behavior when using `terminate_after` in scrolls. For local testing, I have a 1 node 5.6.5 cluster with indices configured with 1 shard: ```auto {:26e1f492-59ca-459a-ab3f-5ed34ee125aa-0 {:settings {:index {:number_of_replicas "0", :search {:slowlog {:threshold {:query {:warn "30s", :trace "10s", :debug "15s", :info "25s"}}}}, :creation_date "1585178759881", :codec "best_compression", :mapper {:dynamic "false"}, :analysis {:filter {:word_delim_with_orig {:type "word_delimiter", :preserve_original "true"}, :possessive_stemmer {:name "possessive_english", :type "stemmer"}}, :analyzer {:default {:filter ["standard" "lowercase" "possessive_stemmer" "kstem"], :tokenizer "classic"}, :sortable {:filter ["lowercase"], :tokenizer "keyword"}, :sonian_standard {:filter ["standard" "lowercase" "possessive_stemmer" "kstem"], :tokenizer "classic"}, :email {:filter ["standard" "lowercase"], :tokenizer "uax_url_email"}}}, :provided_name "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :number_of_shards "1", :uuid "hHJhiISSQPWW9-rrQwtMRQ", :version {:created "5060599"}}}}} ``` I set up my test index with 10 documents that match the following query: ```auto {:sort ["_doc"], :query {:bool {:must {:prefix {:body "foo"}}}}, :_source #{"id"}, :terminate_after 5, :size 2} ``` With a `size = 2` and `terminate_after = 5`, I would expect to get 3 pages back: - 2 2-hit pages and 1 1-hit page. - On the final page, I would also expect `terminated_early = true`. However, what I get back is actually: - 5 2-hit pages, each with `terminated_early = true` and each with a `total = 5` - 1 empty page signifying the scroll has fully "completed" This behavior seems quite odd. Is my understanding of how `size` and `terminate_after` faulty? Does the scroll API honor `terminate_after`? Thanks for the help! Here are the results of the above scroll query, in order: ```auto # --- page 1 {:_scroll_id "DXF1ZXJ5QW5kRmV0Y2gBAAAAAAAAABUWeURLMDFDNmtSRXlOdHV5QzNTcTNNUQ==", :took 1, :timed_out false, :terminated_early true, :_shards {:total 1, :successful 1, :skipped 0, :failed 0}, :hits {:total 5, :max_score nil, :hits [{:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAAAAAAAAAAAAA", :_score nil, :_source {:id "AAAAAAAAAAAAAAAAAAAAAA"}, :sort [0]} {:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAEAAAAAAAAAAQ", :_score nil, :_source {:id "AAAAAAAAAAEAAAAAAAAAAQ"}, :sort [1]}]}} # --- page 2 {:_scroll_id "DXF1ZXJ5QW5kRmV0Y2gBAAAAAAAAABUWeURLMDFDNmtSRXlOdHV5QzNTcTNNUQ==", :took 2, :timed_out false, :terminated_early true, :_shards {:total 1, :successful 1, :skipped 0, :failed 0}, :hits {:total 5, :max_score nil, :hits [{:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAIAAAAAAAAAAg", :_score nil, :_source {:id "AAAAAAAAAAIAAAAAAAAAAg"}, :sort [2]} {:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAMAAAAAAAAAAw", :_score nil, :_source {:id "AAAAAAAAAAMAAAAAAAAAAw"}, :sort [3]}]}} # --- page 3 {:_scroll_id "DXF1ZXJ5QW5kRmV0Y2gBAAAAAAAAABUWeURLMDFDNmtSRXlOdHV5QzNTcTNNUQ==", :took 1, :timed_out false, :terminated_early true, :_shards {:total 1, :successful 1, :skipped 0, :failed 0}, :hits {:total 5, :max_score nil, :hits [{:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAQAAAAAAAAABA", :_score nil, :_source {:id "AAAAAAAAAAQAAAAAAAAABA"}, :sort [4]} {:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAUAAAAAAAAABQ", :_score nil, :_source {:id "AAAAAAAAAAUAAAAAAAAABQ"}, :sort [5]}]}} # --- page 4 {:_scroll_id "DXF1ZXJ5QW5kRmV0Y2gBAAAAAAAAABUWeURLMDFDNmtSRXlOdHV5QzNTcTNNUQ==", :took 1, :timed_out false, :terminated_early true, :_shards {:total 1, :successful 1, :skipped 0, :failed 0}, :hits {:total 5, :max_score nil, :hits [{:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAYAAAAAAAAABg", :_score nil, :_source {:id "AAAAAAAAAAYAAAAAAAAABg"}, :sort [6]} {:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAcAAAAAAAAABw", :_score nil, :_source {:id "AAAAAAAAAAcAAAAAAAAABw"}, :sort [7]}]}} # --- page 5 {:_scroll_id "DXF1ZXJ5QW5kRmV0Y2gBAAAAAAAAABUWeURLMDFDNmtSRXlOdHV5QzNTcTNNUQ==", :took 1, :timed_out false, :terminated_early true, :_shards {:total 1, :successful 1, :skipped 0, :failed 0}, :hits {:total 5, :max_score nil, :hits [{:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAgAAAAAAAAACA", :_score nil, :_source {:id "AAAAAAAAAAgAAAAAAAAACA"}, :sort [8]} {:_index "26e1f492-59ca-459a-ab3f-5ed34ee125aa-0", :_type "msg", :_id "AAAAAAAAAAkAAAAAAAAACQ", :_score nil, :_source {:id "AAAAAAAAAAkAAAAAAAAACQ"}, :sort [9]}]}} # --- page end {:_scroll_id "DXF1ZXJ5QW5kRmV0Y2gBAAAAAAAAABUWeURLMDFDNmtSRXlOdHV5QzNTcTNNUQ==", :took 1, :timed_out false, :terminated_early true, :_shards {:total 1, :successful 1, :skipped 0, :failed 0}, :hits {:total 5, :max_score nil, :hits []}} ``` ---
**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\ **Post date:** [April 22, 2020, 11:47pm UTC](https://discuss.elastic.co/t/5-6-5-terminate-after-and-the-scroll-api/225083/2 "2020-04-22T23:47:54Z")
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.