Solr problem: Could not find configName for collection found:

[#1]

Hi,

I am having a problem with my Solr on Ambari + HDP Stack (Find detailed environment information at the end of question).

My problem is the same as the one in this link.

When I try to create a collection through Solr or create an index through Hue using a csv file, I get the below error:

{ "message": "{\"responseHeader\":{\"status\":400,\"QTime\":16025},\"error\":{\"metadata\":[\"error-class\",\"org.apache.solr.common.SolrException\",\"root-error-class\",\"org.apache.solr.common.cloud.ZooKeeperException\"],\"msg\":\"Error CREATEing SolrCore 'deneme': Unable to create core [deneme] Caused by: Could not find configName for collection deneme found:[twitter_demo, testCollMoney, collection1]\",\"code\":400}}\n (error 400)", "traceback": [ [ "/usr/local/hue/desktop/libs/libsolr/src/libsolr/api.py", 541, "create_core", "response = self._root.post('admin/cores', params=params, contenttype='application/json')" ], [ "/usr/local/hue/desktop/core/src/desktop/lib/rest/resource.py", 132, "post", "allow_redirects=allow_redirects, clear_cookies=clear_cookies)" ], [ "/usr/local/hue/desktop/core/src/desktop/lib/rest/resource.py", 81, "invoke", "clear_cookies=clear_cookies)" ], [ "/usr/local/hue/desktop/core/src/desktop/lib/rest/http_client.py", 173, "execute", "raise self._exc_class(ex)" ] ], "detail": null, "title": "Error while accessing Solr" }

If I try to install examples on Hue, I get the below error:

"responseHeader":{"status":400,"QTime":7},"error":{"metadata":["error-class","org.apache.solr.common.SolrException","root-error-class","java.lang.ClassNotFoundException"],"msg":"Error CREATEing SolrCore 'twitter_demo': Unable to create core [twitter_demo] Caused by: solr.ThaiWordFilterFactory","code":400}}

The only was I can create a collection is uploading zookeeper config first then using ./solr create -c command (both manually on command line). I want to be able to create them over web ui.

Story: First I installed Solr on Ambari while existing Ambari Infra Solr was working. When doing this, Ambari was not starting Solr with /infra-solr Zookeeper path.

Today I removed Solr (keeping Infra-Solr but stopping it) on another host. Result was the same. One tip: even if this installation is a clean one, Solr installation canoeist the configs left from my very first manual Solr installation on tree view.

Environment:

Ambari: 2.4.2

HDP: 2.5.3

Solr: 6.4.0 from https://github.com/abajwa-hw/solr-stack (I changed the repo URL to 6.4.0)

Hue: 3.11 on Docker (Centos 6.8)

Thanks in advance.

[#2]

@Sedat Kestepe

If you want to create collections using the API (https://cwiki.apache.org/confluence/display/solr/Collections+API#CollectionsAPI-CREATE:CreateaCollec...), the configuration (collection.configName) must already be stored in Zookeeper.

[#3]

Hi @Michael Young,

Thank you for your answer.

I see that config must be uploaded first. But that's needed in my installation only. I guess it is not a manual step for expected installation. My question is why my installation needs such a manual step. I mean what kind of configuration is missing?

[#4]

@Sedat Kestepe

This process is not specific to your configuration. The normal way to interact with Solr is to define a schema configuration for a collection and then create the collection using that configuration. The data_driven_schema_config is a schemaless configuraiton. This will allow you to push data to Solr without having to explicitly define every field in advance. However, Solr always requires that you create a collection that is associated with a configuration. The easy approach is to upload the data_driven_schema_config to Zookeeper and create all of your collections using that configuration.

If you use the command tool to create a collection, it will automatically upload the configuration to Zookeeper. If you have a standard config, such as data_driven_schema_configs, you can upload that config to Zookeeper. Then you can use the collection API to create collections multiple times using the same configuration.

You can read more here: https://cwiki.apache.org/confluence/display/solr/Using+ZooKeeper+to+Manage+Configuration+Files

If you just want to upload your configs to Zookeeper, use:

$  bin/solr zk upconfig -n <name for configset> -d <path to directory with configset>

Then you can use the Collection API with the collection.configName parameter to refer to the configuration you uploaded in Zookeeper.

[#5]

Currently command tool also fails with missing config error. I use zookeeper and something is missing on it but I couldn't find. I will review the document you've sent when I find free time. Thank you 🙂