We recently set up a pair of nodes to start ingesting documents (default
setting shards=5), and with about 1 request per minute we're getting
timeouts in some POST requests. This means that elasticsearch is not
replying in 5 minutes time. In paramedic we see though that the indexing
time is about 32 secs (which is high anyway compared to what we were
expecting, but is way lower than 5 minutes! which is the default time out
of the client-stack that makes the Http request).
We're using version 0.19.9. Does this ring a bell to anyone? Any bug that
could have been fixed recently in newer versions which was affecting this
one?
Which client are you using? Are you bulk indexing or sending one index
request per document? What is the size of your document/total size if bulk?
On Tue, Nov 13, 2012 at 8:32 AM, Andrés .G Aragoneses knocte@gmail.comwrote:
Hello!
We recently set up a pair of nodes to start ingesting documents (default
setting shards=5), and with about 1 request per minute we're getting
timeouts in some POST requests. This means that elasticsearch is not
replying in 5 minutes time. In paramedic we see though that the indexing
time is about 32 secs (which is high anyway compared to what we were
expecting, but is way lower than 5 minutes! which is the default time out
of the client-stack that makes the Http request).
We're using version 0.19.9. Does this ring a bell to anyone? Any bug that
could have been fixed recently in newer versions which was affecting this
one?
Size of each document is minimal, small JSON object that more or less fits
in a screen.
Right now the total number of documents in the cluster is 14,200 (26 MB).
(BTW I made a typo in my first message: we're obviously using PUT, not POST)
Thanks for your reply!
On 13 November 2012 17:27, Ivan Brusic ivan@brusic.com wrote:
Many questions:
Which client are you using? Are you bulk indexing or sending one index
request per document? What is the size of your document/total size if bulk?
On Tue, Nov 13, 2012 at 8:32 AM, Andrés .G Aragoneses knocte@gmail.comwrote:
Hello!
We recently set up a pair of nodes to start ingesting documents (default
setting shards=5), and with about 1 request per minute we're getting
timeouts in some POST requests. This means that elasticsearch is not
replying in 5 minutes time. In paramedic we see though that the indexing
time is about 32 secs (which is high anyway compared to what we were
expecting, but is way lower than 5 minutes! which is the default time out
of the client-stack that makes the Http request).
We're using version 0.19.9. Does this ring a bell to anyone? Any bug that
could have been fixed recently in newer versions which was affecting this
one?
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.