When the data you want to access is not stored in COS, you can use COS’s origin-pull feature to pull data from a specified origin server (e.g., a local IDC, or the origin server/bucket of other cloud vendors).
Origin-pull is mainly used for hot data migration, redirection for specified requests, and other scenarios. You can configure it as needed. For detailed directions, please see Setting Origin-Pull.
After an origin-pull rule is set, if the requested object does not exist in the bucket, the object can be found from the configured origin server address using the origin-pull rule and returned to the user. Likewise, when you need to redirect specific requests, the origin-pull rule can be used for COS to access data in the origin server.
If the origin-pull is asynchronous, after origin-pull is configured, 302 will be returned for the first COS request, and the client’s second request will be forwarded to the origin server. If the origin-pull is synchronous, COS will pull data from the origin server in real time and return it to the client, and also dump the data to the server.
SLA of the offline origin-pull module is not always successful. If you want to know whether the origin-pull upload is successful, you can go to the SCF console to set a callback that is triggered by the offline origin-pull. For more information about SCF’s COS triggers, please see COS Trigger.
An origin-pull address is usually an IP or a domain name. It specifies where the data you need to pull is stored. When COS does not have the resource you want to access, you can use the origin-pull address to pull the resource in real time.
Yes. COS will pull the resource automatically and create a path.