Backup repository missing exception

Elastic Stack Elasticsearch
1

Hi Elasticsearch Team,

After redeploying Elasticsearch 1.7 in our server, data has been removed. and when we use _snapshot api to restore created backup using:
curl -X GET "localhost:9200/_snapshot/test_backup"

it shows the following exception:

`{"error":"RepositoryMissingException[[test_backup] missing]","status":404}`

the request:

curl -X GET "localhost:9200/_snapshot

Has an empty response

{}

Please help me.

Thank you!

2

You need to re-register the repository with the same settings that your used to create snapshots.

3

Hi @Igor_Motov,

Thank you for the feedback!
But i have already added the path in elasticsearch_cutom.yml as repo path, and the feedback was already created by an other instance of elasticsearch.

Best regards,
Mohammed ESSABRI

4

Your errors indicate the repository registration didn't succeed.

5

Thank you @Igor_Motov for your help!

I will try with that.

Best regards,

6

Dear @Igor_Motov,

I've created a snapshot named snapshot20180625 in an environment and i copied it into another environment, then i tried to register the repository using:

curl -XPUT "***IP***:9200/_snapshot/snapshot20180625" -d '{
  "type": "fs",
  "settings": {
    "compress": "true",
    "location": "/shared/appdata/sls/sls.elasticsearch/backups"
  }
}'

But it gives me the following error:

{"error":"RemoteTransportException[[***server name***][inet[/***IP***:9300]][cluster:admin/repository/put]]; nested: ProcessClusterEventTimeoutException[failed to process cluster event (put_repository [snapshot20180625]) within 30s]; ","status":503}

Can you please give me a help ?
Thank you!

7

There is one to many relationship between repositories and snapshots and you seemed to be conflating the two. One repository can contain multiple snapshots and you register repository while creating and restoring snapshots. Please, check documentation and fully understand the difference and relationship between them before continue.

The error that you are getting can indicate several things. It can indicate that some nodes don't have access to the directory /shared/appdata/sls/sls.elasticsearch/backups and timeout trying to access it or that the cluster is in pretty bad shape and cannot process the request within 30 seconds.

Could you verify that the directory /shared/appdata/sls/sls.elasticsearch/backups is mounted on all nodes and have the same access rights and that elasticsearch is running under the user with the same id on all nodes.

Could you also check that nothing else is going on on the cluster by running GET /_cat/pending_tasks

Do you see any error message in the log file on the master node?

1 Like
8

Dear @Igor_Motov,

I would thank you for the help, it was exactly as you suggested, the owner user was not the same for both repository and Elasticsearch cluster.

Thank you so much!

Best regards,
Mohammed ESSABRI

9

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.