Set the number of connections to be used for upload or download.
Using S3 Storage
When you use the Warp engine with -XX:CRaCEngine=warp, the checkpoint can be stored on S3-compatible buckets. Define the location starting with s3://, for example, -XX:CRaCCheckpointTo=s3://bucket/some-key.
You must configure the environment with credentials in ~/.aws or environment variables, as specified by the standard approach of the AWS SDK. You can finetune some settings with -XX:CRaCEngineOptions:
| Setting | Description | Default |
|---|---|---|
|
|
|
|
The endpoint by default is configured to upload to AWS S3. Use this setting to change the endpoint URL and upload to another S3 compatible storage. The endpoint must start with |
AWS S3 |
|
Configure AWS logging level. |
Off |
|
Configure AWS logging prefix. |
|
|
Images stored in S3 can be split over several upload connections. AWS supports up to 10000 parts each between 5MB and 5GB. When your checkpoint is larger than 50GB, you might need to increase the value. |
5MB |
|
Images stored in S3 can be split over several download connections and the size can be specified with this setting. |
5MB |
|
Note
|
All available settings for -XX:CRaCEngineOptions=<value> can be listed by starting your JVM with -XX:CRaCEngine=warp -XX:CRaCEngineOptions=help.
|