Granted Patent B2

US 11,615,028 · App. 17/237,359 · Granted Mar 28, 2023

Inventors: Jenny Derzhavetz (Raanana, IL); Vladimir Shveidel (Pardes-Hana, IL); Dror Zalstein (Givatayim, IL); Bar David (Rishon Lezion, IL)

Assignee: EMC IP Holding Company, LLC

G06F12/0882G06F9/30047G06F12/0238G06F12/123

Abstract

Claims (49)

1. A computer-implemented method, executed on a computing device, comprising:

receiving a flush request for a metadata page stored in a storage array of a multi-node storage system, wherein the flush request is a request to commit the metadata page from a cache memory system to the storage array, thus defining a commit request;

queueing the flush request on a flush request lock queue on at least one node of the multi-node storage system; and

processing, via multiple nodes of the multi-node storage system, one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue, wherein processing, via the multiple nodes of the multi-node storage system, the one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue includes:

determining, at each node of the multi-node storage system, whether a destage request for the metadata page is being processed on any node of the multi-node storage system based upon, at least in part, the flush request lock queue;

in response to determining that the destage request for the metadata page is being processed on at least one node, waiting until the destage request is cleared from the flush request lock queue; and

in response to determining that the destage request for the metadata page is not being processed on any node, processing, via one node of the multi-node storage system, the commit request.

2. The computer-implemented method of claim 1 , wherein the flush request is a request to destage the metadata page to the storage array, thus defining a destage request.

3. The computer implemented method of claim 2 , wherein processing, via the multiple nodes of the multi-node storage system, the one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue includes:

reading the metadata page from the storage array;

determining a current tablet identifier for the metadata page; and

processing the destage request based upon, at least in part, the current tablet identifier for the metadata page.

4. The computer implemented method of claim 3 , wherein processing the destage request based upon, at least in part, the current tablet identifier for the metadata page includes:

skipping the destage request when the current tablet identifier for the metadata page indicates that the metadata page read from the storage array is newer than the metadata data page to destage; and

processing the destage request when the current tablet identifier for the metadata page indicates that the metadata page read from the storage array is older than the metadata data page to destage.

5. The computer implemented method of claim 1 , further comprising:

determining that at least one node of the multi-node storage system has failed;

pausing flush requests for each metadata page referenced within the flush request lock queue; and

flushing each metadata page referenced within the flush request lock queue from a non-volatile storage area to the storage array.

6. A computer program product residing on a non-transitory computer readable medium having a plurality of instructions stored thereon which, when executed by a processor, cause the processor to perform operations comprising:

receiving a flush request for a metadata page stored in a storage array of a multi-node storage system, wherein the flush request is a request to destage the metadata page to the storage array, thus defining a destage request;

queueing the flush request on a flush request lock queue on at least one node of the multi-node storage system; and

processing, via multiple nodes of the multi-node storage system, one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue, wherein processing, via the multiple nodes of the multi-node storage system, the one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue includes:

reading the metadata page from the storage array;

determining a current tablet identifier for the metadata page; and

processing the destage request based upon, at least in part, the current tablet identifier for the metadata page.

7. The computer program product of claim 6 , wherein the flush request is a request to commit the metadata page from a cache memory system to the storage array, thus defining a commit request.

8. The computer program product of claim 7 , wherein processing, via the multiple nodes of the multi-node storage system, the one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue includes:

determining, at each node of the multi-node storage system, whether a destage request for the metadata page is being processed on the node based upon, at least in part, the flush request lock queue;

in response to determining that the destage request for the metadata page is being processed on at least one node, waiting until the destage request is cleared from the flush request lock queue; and

in response to determining that the destage request for the metadata page is not being processed on any node, processing, via one node of the multi-node storage system, the commit request.

9. The computer program product of claim 6 , wherein processing the destage request based upon, at least in part, the current tablet identifier for the metadata page includes:

skipping the destage request when the current tablet identifier for the metadata page indicates that the metadata page read from the storage array is newer than the metadata data page to destage; and

processing the destage request when the current tablet identifier for the metadata page indicates that the metadata page read from the storage array is older than the metadata data page to destage.

10. The computer program product of claim 6 , wherein the operations further comprise:

determining that at least one node of the multi-node storage system has failed;

pausing flush requests for each metadata page referenced within the flush request lock queue; and

flushing each metadata page referenced within the flush request lock queue from a non-volatile storage area to the storage array.

11. A computing system comprising:

a memory; and

a processor configured to receive a flush request for a metadata page stored in a storage array of a multi-node storage system, wherein the flush request is a request to destage the metadata page to the storage array, thus defining a destage request, wherein the processor is further configured to queue the flush request on a flush request lock queue on at least one node of the multi-node storage system, and wherein the processor is further configured to process, via multiple nodes of the multi-node storage system, one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue, wherein processing, via the multiple nodes of the multi-node storage system, the one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue includes: reading the metadata page from the storage array, determining a current tablet identifier for the metadata page, and processing the destage request based upon, at least in part, the current tablet identifier for the metadata page.

12. The computing system of claim 11 , wherein the flush request is a request to commit the metadata page from a cache memory system to the storage array, thus defining a commit request.

13. The computing system of claim 12 , wherein processing, via the multiple nodes of the multi-node storage system, the one or more flush requests on the metadata page based upon, at least in part, the flush request lock queue includes:

determining, at each node of the multi-node storage system, whether a destage request for the metadata page is being processed on the node based upon, at least in part, the flush request lock queue;

in response to determining that the destage request for the metadata page is being processed on at least one node, waiting until the destage request is cleared from the flush request lock queue; and

in response to determining that the destage request for the metadata page is not being processed on any node, processing, via one node of the multi-node storage system, the commit request.

14. The computing system of claim 11 , wherein processing the destage request based upon, at least in part, the current tablet identifier for the metadata page includes:

skipping the destage request when the current tablet identifier for the metadata page indicates that the metadata page read from the storage array is newer than the metadata data page to destage; and

processing the destage request when the current tablet identifier for the metadata page indicates that the metadata page read from the storage array is older than the metadata data page to destage.

Assignments (10)