Any point in time backups and tier one storage replication
Generating any point in time backups from secondary storage without native snapshot generation and providing failover capabilities from a primary storage. Data or IOs from a source are distributed to both the primary storage and the secondary storage. When a disaster occurs with one of these storages, recovery of one of the storages can be achieved using a delta marker and data from the other of the storages.
1. A method, comprising:
receiving an IO stream from a source;
preparing to distribute the IO stream to a first target as a first IO stream and a second target as a second IO stream, the IO stream including a plurality of IOs;
sending the first IO stream to the first target in a first manner and sending the second IO stream to the second target in a second manner that is different from the first manner;
storing a delta marker that identifies whether the plurality of IOs have been distributed to both the first target and the second target;
detecting an issue with the first target such that distribution to the first target is affected; and
resolving the issue using the delta marker and data stored at the second target,
wherein the first target is one of a primary storage and a secondary storage and the second target is the other of the primary storage and the secondary storage,
wherein when the first target is the primary storage, the first manner occurs without accumulating the first IO stream for large transfer and when the second target is the primary storage, the second manner occurs without accumulating the second IO stream for large transfer.
2. The method of claim 1 , further comprising resolving the issue without disrupting the source for the data.
3. The method of claim 1 , further comprising continuing to distribute IOs from the source to the second target while resolving the issue with the first target.
4. The method of claim 1 , wherein the first target is a primary storage and the second target is a secondary storage, the method comprising recovering the primary storage from the second storage by:
constructing synchronization metadata from a metadata stream on the second storage and from the delta marker;
reading data from the second storage based on the synchronization metadata; and
applying the read data to the primary storage.
5. The method of claim 3 , further comprising receiving a new IO from the source and, while recovering the first target from the second target:
applying the new IO normally when the new IO does not apply to areas associated with the synchronization metadata;
applying the new IO normally when the new IO applies to an area of the primary storage that has already been synchronized; or
applying the new IO when the new IO applies to an area of the primary that is subject to synchronization but has not been synchronized and deleting IOs from the synchronization metadata related to the area associated with the new IO.
6. The method of claim 1 , wherein the first target is a secondary storage and the second target is the primary storage, the method further comprising recovering the secondary storage from the primary storage.
7. The method of claim 6 , further comprising:
reading data from the primary storage based on the delta marker;
storing new IOs from the source in a synchronization stream;
applying the read data to the secondary storage; and
applying the synchronization stream after the secondary storage is recovered from the primary storage.
8. The method of claim 1 , further comprising generating a PiT image from the secondary storage.
9. The method of claim 1 , further comprising failing over to a latest copy stored on the primary storage.
10. The method of claim 1 , further comprising failing over to a near-latest copy on the primary storage by writing data from the secondary storage to the primary storage by identifying locations on the primary storage to be overwritten and writing data from the secondary storage that corresponds to the locations and a selected point in time.
11. A non-transitory storage medium having stored therein instructions that are executable by one or more hardware processors to perform operations comprising:
receiving an IO stream from a source;
preparing the IO stream for distribution to a first target as a first IO stream and a second target as a second IO stream, the IO stream including a plurality of IOs;
sending the first IO stream to the first target in a first manner without accumulating for large data transfer and sending the second IO stream to the second target in a second manner that is different from the first manner;
storing a delta marker that identifies whether the plurality of IOs have been distributed to both the first target and the second target;
detecting an issue with the first target such that distribution to the first target is affected; and
resolving the issue using the delta marker and data stored at the second target.
12. The non-transitory storage medium of claim 11 , wherein the first target is one of a primary storage and a secondary storage and, the second target is the other of the primary storage and the secondary storage.
13. The non-transitory storage medium of claim 12 , further comprising continuing to distribute IOs from the source to the second target while resolving the issue with the first target.
14. The non-transitory storage medium of claim 11 , wherein the first target is a primary storage and the second target is a secondary storage, the method comprising recovering the primary storage from the second storage by:
constructing synchronization metadata from a metadata stream on the second storage and from the delta marker;
reading data from the second storage based on the synchronization metadata; and
applying the read data to the primary storage.
15. The non-transitory storage medium of claim 14 , further comprising receiving a new IO from the source and, while recovering the first target from the second target:
applying the new IO normally when the new IO does not apply to areas associated with the synchronization metadata;
applying the new IO normally when the new IO applies to an area of the primary storage that has already been synchronized; or
applying the new IO when the new IO applies to an area of the primary that is subject to synchronization but has not been synchronized and deleting IOs from the synchronization metadata related to the area associated with the new IO.
16. The non-transitory storage medium of claim 11 , wherein the first target is a secondary storage and the second target is the primary storage, the method further comprising recovering the secondary storage from the primary storage.
17. The non-transitory storage medium of claim 16 , further comprising:
reading data from the primary storage based on the delta marker;
storing new IOs from the source in a synchronization stream;
applying the read data to the secondary storage; and
applying the synchronization stream after the secondary storage is recovered from the primary storage.
18. The non-transitory storage medium of claim 11 , further comprising generating a PiT image from the secondary storage.
19. The non-transitory storage medium of claim 11 , further comprising failing over to a latest copy stored on the primary storage.
20. The non-transitory storage medium of claim 11 , further comprising failing over to a near-latest copy on the primary storage by writing data from the secondary storage to the primary storage by identifying locations on the primary storage to be overwritten and writing data from the secondary storage that corresponds to the locations and a selected point in time.