We are having a process feeding the logs into our ES server (single
node, 5 shards, vanilla setup). One thing puzzles us is that the
max_doc, deleted_docs and size number are jumping up and down in "ES
head" (by >1000 documents).
The num_docs appears OK though, and we are pretty sure no delete
operation is there.
If you are reindexing the same doc, it ends up deleting the previous doc and
indexing the new doc. Also, as internal merges happen within the index to
maintain it, those number may change as well.
We are having a process feeding the logs into our ES server (single
node, 5 shards, vanilla setup). One thing puzzles us is that the
max_doc, deleted_docs and size number are jumping up and down in "ES
head" (by >1000 documents).
The num_docs appears OK though, and we are pretty sure no delete
operation is there.
I had noticed this in my setup and problem turned out that I was inserting records with duplicate ids.
I was not noticing numbers jump up and down though, and total of numdocs and deleted docs was equal to the number of docs that i had inserted...
-----Original message-----
From: Liyu
Sent: 28/10/2011, 12:35 am
To: elasticsearch
Subject: max_doc, deleted_docs and size number jump up and down
Hi,
We are having a process feeding the logs into our ES server (single
node, 5 shards, vanilla setup). One thing puzzles us is that the
max_doc, deleted_docs and size number are jumping up and down in "ES
head" (by >1000 documents).
The num_docs appears OK though, and we are pretty sure no delete
operation is there.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.