The S3 Backup Restore Tool is the standalone utility for performing DR from your S3 backup bucket, either to the original cluster or to an empty cluster that is meant to replace the original. See S3 Backup Feeds.
Once your data is backed up in S3, the restore tool allows both examining a backup and control how, what, and where it is restored:
...
| Section | Settings |
|---|
[s3] | - host — The hostname of the S3 service.
- port — The port to use for the S3 service. Use 443 or else 80, if SSL (sslOption) is disabled.
- accessKeyID — The S3 access key ID.
- secretAccessKey — The S3 secret access key.
- bucketName — The name of the destination bucket in S3.
- sslOption — The S3 connection constraint, with one of two values:
- "trusted" (the default) specifies use of SSL and requires a trusted server certificate from the destination server.
- “none” disables use of SSL. Use only for testing and troubleshooting, and change the port to 80.
|
|---|
[s3] archival only | If you are using an S3 bucket with an archival storage class (Glacier, Glacier Deep Archive), set these additional parameters: performArchiveRetrieval — Whether restoration from archival storage is needed. If false (default), performing a restore will not incur any expenses for the bucket owner. retrievalTier — Which S3 Glacier retrieval tier to use for restoration: ‘Standard' (default), 'Expedited', or 'Bulk'. Each tier has its own cost and expected restoration time; see Amazon S3 Storage Classes. accountID — Specifies the 9-digit AWS account ID of the bucket owner, granting the tool permission to incur archive restoration expenses at the tier requested. This setting appears in the x-amz-expected-bucket-owner header on the restore object request. activeLifetimeDays — How many days an object restored from archive should remain active before expiring (returning to archival storage). The default is 7 (1 week).
|
|---|
[forwardProxy] | This section is for use only with an optional forward proxy: - host — The forward proxy hostname or IP address.
- port — The forward proxy host to use.
- username — (optional) The user name.
- password — (optional) The password.
|
|---|
[log] | You may use the same log settings as your Swarm cluster; if you do so, identify the logs by looking for those with the component "RESTORE". - host — The log host. Leave blank to disable logging.
- port — (optional) The log port. Defaults to 514.
- file — (optional) The log filename. Accepts the value of “stdout” for logging to the console screen. Defaults to
/dev/null. - level — The log level. Defaults to 30 (Warning). Levels are the same used by Swarm: 20 (Info), 15 (Audit), 10 (Debug).
|
|---|
[swarm] | - host — A list of host names or IP addresses of Swarm nodes or Gateway nodes.
- port — (optional) The SCSP port. Defaults to 80.
- user — The cluster administrator user name, usually "
admin". - password — The cluster administrator password.
|
|---|
Additional Restore Configuration
- Gateway — If you are communicating with Swarm cluster via Gateway, be sure to add Add the IP of the machine where the Restore tool will run to the Gateway configuration setting
scsp.allowSwarmAdminIP if communicating with a Swarm cluster via Gateway.
Using the Restore Tool
| Info |
|---|
| title | Full cluster restore |
|---|
|
Before undertaking a restore of a large cluster, contact DataCore Support. They will help you balance the speed of your restore with your bandwidth constraints by examining the space used by the S3 backup bucket, estimating the bandwidth needed, and recommending best use of the -p command line option (for multiple simultaneously running restore commands on different hosts). They will also advise you on whether you need a forward proxy, to reduce bandwidth usage. If you are using an AWS Glacier storage class, you might pull your AWS bucket out of cold storage before your full cluster restore by changing the storage class to Standard. |
...
