Cloud Search troubleshooting

Available in Classic and VPC

You might run into the following problems when using Cloud Search Find out causes and possible solutions.

Container error

Search container shows an error state.

Cause

The search container’s memory usage grows with query traffic and document volume and the indexing container’s memory usage grows with document volume, number of indexes, and index complexity. When search-container memory usage exceeds 80%, or when memory usage is high relative to the container type in use, the console marks the container as Error.

Solution

Replace the container or upgrade the container type.

"Document format is wrong" error message

You get "Document format is wrong" when uploading documents in bulk.

Cause

• There is an issue with the document format you are uploading.
• The value specified in docid differs from the key value.
• You are requesting a section that does not exist.
• For document uploads, the value of content must be of type String.

Solution

• The upload format must be *.json. Check the document format.
• The section value designated by key must match the key field value. Verify the specified values.
• Check whether you are requesting a non-existent section.
• For document uploads, write the content value as a String. If the value is null, use an empty string "", send numbers as strings.

Synonyms and typo correction

You cannot find options for synonyms and typo correction.

Cause

Synonyms and typo correction are not supported.

Solution

As explained in the cause section, it is a normal operation and it does not need any action.

N-gram analysis

I want to enable Ngram analysis.
I want to enable N-gram search in results in addition to a thesaurus or manual indexing.

Cause

You cannot edit indexes in the console, and you can use the API to extract ngram terms as index tokens.

Solution

See the Schema update API guide and apply as follows:

...
   {
     "type": "hanaterm",
     "method": "ngram",
     "option": "ngram=3 +ngramonly" # ngram option: Extracts tokens from unigrams up to ngram as index terms. (Maximum and default are 10; minimum is 2) The ngramonly option extracts only ngrams of the specified size as index terms.
   }

...

Cannot edit indexes

You cannot modify an index.

Cause

You cannot modify indexes in the console. Index modifications are available only via the API.

Solution

As explained in the cause section, it is a normal operation and it does not need any action. For how to modify an index, see the Schema update API guide.

Learning resources

We offer various materials for you to explore. To learn more about Cloud Search, check out these helpful links:

• FAQs: Cloud Search FAQs
• API guide: Cloud Search API guide
• NAVER Cloud Platform blog: Build your search environment easily with Cloud Search