Hi, I work with the OP and have been digging into this. I don't have a
hard example to send you just yet, but have uncovered some important
details. We use a parent/child mapping to index attachments separate
from our metadata and it appears to only occur when doing a
"match_all" against both the parent (content) and child type
(contentFiles), eg:
curl -XGET http://localhost:9200/temp/_search -d '{"query":
{"match_all":{}}}'
curl -XGET http://localhost:9200/temp/content,contentFiles/_search -d
'{"query":{"match_all":{}}}'
both fail, but:
curl -XGET http://localhost:9200/temp/content/_search -d '{"query":
{"match_all":{}}}'
curl -XGET http://localhost:9200/temp/contentFiles/_search -d
'{"query":{"match_all":{}}}'
do not. Here is log snippet, showing the index being dropped, created,
and several mapping updates before the exception trace:
[2011-07-12 18:57:14,692][INFO ][cluster.metadata ] [Nital,
Adri] [temp] deleting index
[2011-07-12 18:57:16,879][INFO ][cluster.metadata ] [Nital,
Adri] [temp] creating index, cause [auto(index api)], shards [5]/[1],
mappings [content, contentFiles]
[2011-07-12 18:57:16,927][INFO ][cluster.metadata ] [Nital,
Adri] [temp] update_mapping [content] (dynamic)
[2011-07-12 18:57:17,122][INFO ][cluster.metadata ] [Nital,
Adri] [temp] update_mapping [content] (dynamic)
[2011-07-12 18:57:17,244][INFO ][cluster.metadata ] [Nital,
Adri] [temp] update_mapping [content] (dynamic)
[2011-07-12 18:57:17,372][INFO ][cluster.metadata ] [Nital,
Adri] [temp] update_mapping [content] (dynamic)
[2011-07-12 18:57:17,700][INFO ][cluster.metadata ] [Nital,
Adri] [temp] update_mapping [content] (dynamic)
[2011-07-12 18:57:18,122][DEBUG][action.search.type ] [Nital,
Adri] [257386] Failed to execute fetch phase
org.elasticsearch.search.fetch.FetchPhaseExecutionException: [temp]
[0]: query[ConstantScore(:)],from[0],size[10]: Fetch Failed [Failed
to load uid from the index]
at
org.elasticsearch.search.fetch.FetchPhase.extractUid(FetchPhase.java:
167)
at org.elasticsearch.search.fetch.FetchPhase.execute(FetchPhase.java:
80)
at
org.elasticsearch.search.SearchService.executeFetchPhase(SearchService.java:
352)
at
org.elasticsearch.search.action.SearchServiceTransportAction.sendExecuteFetch(SearchServiceTransportAction.java:
314)
at
org.elasticsearch.action.search.type.TransportSearchQueryThenFetchAction
$AsyncAction.executeFetch(TransportSearchQueryThenFetchAction.java:
140)
at
org.elasticsearch.action.search.type.TransportSearchQueryThenFetchAction
$AsyncAction.access$200(TransportSearchQueryThenFetchAction.java:59)
at
org.elasticsearch.action.search.type.TransportSearchQueryThenFetchAction
$AsyncAction$1.run(TransportSearchQueryThenFetchAction.java:113)
at java.util.concurrent.ThreadPoolExecutor
$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor
$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:662)
Our documents are rather complex so I'm not exactly sure what about
that particular document (if anything) triggered the issue. Other
child documents were inserted by my test script but did not cause any
shard failures for the match_all query. Unfortunately I can't post the
data so I will try to whittle it down to something simple you will be
able to reproduce on your end.
On Jul 7, 5:58 pm, Shay Banon shay.ba...@elasticsearch.com wrote:
It seems like its missing the internal _uid field that is stored in each document from the document fetched. Is it something reproducible? (I improved the failure message to include the loaded fields).
On Friday, July 8, 2011 at 12:10 AM, lmader wrote:
Failed to
load uid from the index