unable to take solr backup using s3

[kannanvr]

Hi Team,

Trying to take Solr backup to s3. Getting below error.

some directory are created in s3, but no data in it.

s3://solr/solrbkp-country/country/
[2023-08-02 22:06:29 IST]     0B index/77b5e96f-c787-449d-86c3-d15a3dd96fa2
[2023-08-02 22:06:29 IST]     0B shard_backup_metadata/
[2023-08-02 22:06:29 IST]     0B zk_backup_0/

Kindly refer the solrCloud.spec for backuprepositories and solrBackup yaml (Also collection country was created and inserted few data as well).

  backupRepositories:
  - name: "s3-backup"
    s3:
      baseLocation: "/"
      bucket: "solr"
      endpoint: "http://x.x.x.x"
      region: "us-east-1"
      credentials:
        accessKeyIdSecret:
          name: s3-cred
          key: access_key
        secretAccessKeySecret:
          name: s3-cred
          key: secret_key
---
apiVersion: solr.apache.org/v1beta1
kind: SolrBackup
metadata:
  name: solrbkp
  namespace: solr-cluster-2
spec:
  repositoryName: "s3-backup"
  solrCloud: test1
  collections:
    - country
  recurrence: # Store one backup daily, and keep a week at a time.
    schedule: "@every 5m"
    maxSaved: 7

[HoustonPutman]

I see the following error "Parent directory doesn't exist". That's a very strange error, and I think it's an error in the S3-repository plugin (or the backupManager itself), not the Solr Operator...

Is this happen everytime you try to backup or did it happen just once?

[kannanvr]

Hi @HoustonPutman ,

We are getting this error every-time. As i mentioned, the directory structure is created in s3, but files are not copied.

[HoustonPutman]

Yes, please bring this up on the Solr users mailing list, as it's a Solr issue not a Solr Operator issue. Sorry for the delay.

As a guess, I think you are using an "S3 compatible" product that Solr doesn't actually support, such as Minio. If that's the case, I'm not sure there will be a solution.