[ElasticSearch indexation] How to bypass : "Ignore indexing command in bulk, blob does not exists anymore"

Hi,

I use :

  • Nuxeo 10.10
  • ES : 6.5.4

I try to move documents in ES, but i have no blob file (because this is a POC and i have only the postgresql database). So when i try to indexing a file type document i can see in log (elastic.log)

2019-03-01T14:57:01,446 INFO [Nuxeo-Work-elasticSearchIndexing-9:152939146812607.226469385] [org.nuxeo.elasticsearch.core.ElasticSearchIndexingImpl] Ignore indexing command in bulk, blob does not exists anymore: {"id":"IxCd-17414654","type":"INSERT","docId":"cc8d2f15-7f49-47ae-9db7-edc6ff4ac096","path":"/default-domain/workspaces/XXXXXX/22144415","repo":"default","recurse":false,"sync":false,"order":1551448614340426}

How can i bypass this ? I need to indexing the document even if there are no blob file.

Thank you.

0 votes

1 answers

73 views

ANSWER



Here is what i do to bypass the problem.

  • I created an empty text file

  • I hash it (hash = cfcd208495d565ef66e7dff9f98764da)

  • I stored the binary file in default repository (filesystem)

    data\binaries\data\cf\cd\cfcd208495d565ef66e7dff9f98764da

  • I created a backup table create table content_backup as select id,data from content

  • I updated the table 'content' according to the hash.

update content set data = 'cfcd208495d565ef66e7dff9f98764da'

Now Elasticsearch find a 'fake' binary document and it can index my documents (only metadata / schema of course).

0 votes