It is probably a Zookeeper limit. You have to set jute.maxbuffer in the Java System properties of all (!) zookeeper Servers and clients to the same value (in your case it should be a little bit larger than your largest file). If possible you can try to avoid storing the NLP / ML models in Solr but provide them on a share or similar where all Solr nodes have access to.
> Am 11.06.2019 um 00:32 schrieb Pratik Patel <pra...@semandex.net>: > > Hi, > > I am trying to upload a configuration to "MiniSolrCloudCluster" in my unit > test. This configuration has some binary files for NLP related > functionality. Some of these binary files are bigger than 5 MB. If I try to > upload configuration with these files then it doesn't work. I can set up > the cluster fine if I remove all binary files bigger than 5 MB. > > I have noticed the same issue when I try to restore a backup having > configuration files bigger than 5 MB. > > Does jetty have some limit on the size of configuration files? Is there a > way to override this? > > Thanks, > Pratik
