CONTINUOUS DATA PROTECTION IN CLOUD USING STREAMS
One example method includes performing a recovery operation. A recovery operation is performed using streams rather than volumes in the cloud. Do data is written to a do stream. When the do data is applied to a cloud volume, which is a copy or replica of a production volume, data being replaced is saved to an undo stream to ensure that any PiT can be recovered using one or more of the cloud volume, the do stream and/or the undo stream.
1 . A method, comprising:
receiving writes from a data protection system configured to provide data protection operations to a production site into a do stream in a cloud, the writes including data;
instantiating a compute instance to read data from the do stream;
moving data to be overwritten with the data from the do stream from a cloud volume to an undo stream by the compute instance; and
writing the data read from the do stream to the cloud volume by the compute instance.
2 . The method of claim 1 , wherein the do stream comprises a queue and the undo stream comprises a queue.
3 . The method of claim 1 , further comprising performing a recovery operation to recovery a selected point-in-time (PiT) backup.
4 . The method of claim 3 , further comprising determining whether the selected PiT backup has been applied to the cloud volume.
5 . The method of claim 4 , further comprising, when the selected PiT backup has been applied to the cloud volume, recovering the selected PiT backup from the cloud volume and the undo stream.
6 . The method of claim 5 , further comprising performing a snapshot of the selected PiT once the cloud volume is restored to the selected PiT.
7 . The method of claim 4 , further comprising reading from the do stream when the selected PiT backup has not been applied to the cloud volume, and applying the read data to the cloud volume.
8 . The method of claim 7 , further comprising, writing data on the cloud volume being replaced to the undo stream.
9 . The method of claim 3 , further comprising continuing to receive new data into the do stream while performing the recovery operation.
10 . The method of claim 1 , further comprising occasionally starting up the compute instance.
11 . A non-transitory storage medium having stored therein instructions that are executable by one or more hardware processors to perform operations comprising:
receiving writes from a data protection system configured to provide data protection operations to a production site into a do stream in a cloud, the writes including data;
instantiating a compute instance to read data from the do stream;
moving data to be overwritten with the data from the do stream from a cloud volume to an undo stream by the compute instance; and
writing the data read from the do stream to the cloud volume by the compute instance.
12 . The non-transitory storage medium of claim 11 , wherein the do stream comprises a queue and the undo stream comprises queue.
13 . The non-transitory storage medium of claim 11 , further comprising performing a recovery operation to recovery a selected point-in-time (PiT) backup.
14 . The non-transitory storage medium of claim 13 , further comprising determining whether the selected PiT backup has been applied to the cloud volume.
15 . The non-transitory storage medium of claim 14 , further comprising, when the selected PiT backup has been applied to the cloud volume, recovering the selected PiT backup from the cloud volume and the undo stream.
16 . The non-transitory storage medium of claim 15 , further comprising performing a snapshot of the selected PiT once the cloud volume is restored to the selected PiT.
17 . The non-transitory storage medium of claim 14 , further comprising reading from the do stream when the selected PiT backup has not been applied to the cloud volume, and applying the read data to the cloud volume.
18 . The non-transitory storage medium of claim 17 , further comprising, writing data on the cloud volume being replaced to the undo stream.
19 . The non-transitory storage medium of claim 13 , further comprising continuing to receive new data into the do stream while performing the recovery operation.
20 . The non-transitory storage medium of claim 11 , further comprising occasionally starting up the compute instance.