Hi,
following comment is mentioned in
https://github.com/apache/lucene-solr/blob/master/lucene/core/src/java/org/apache/lucene/util/fst/package-info.java
.
"Input values (keys). These must be provided to Builder in Unicode code
point (UTF8 or UTF32) sorted order. Note that sorting by Java's
Str
Years ago, working on the Ultraseek spider, we did a bunch of tests on
compressed HTTP.
I expected it to be a big win, but the results were really inconclusive.
Sometimes it was faster,
sometimes it was slower. We left it turned off.
It is an absolute win for serving already-compressed static co
We have 12 node SOLR cloud with 3 zookeeper ensemble
RAM: 80 CPU:40 Heap:16GB Records: 4 Million
We do real time update and deletes (by ID), and we do us Inplace updates
for 4 fields
We have one index with 4 shards: 1 shard in 3 nodes
Often we are getting the following errors
1. *2021-01-08 17:
You're right Matthew.
Jetty supports it for responses but for requests it doesn't seem to be the
default.
However I found a configuration not documented that needs to be set in the
GzipHandler for it to work: inflateBufferSize.
For SolrJ it still hacky to send gzip requests, maybe easier to use