A method, device and medium for processing file gateway failure switching

Abstract

The present invention discloses a processing method, device and medium for file gateway fault switching. When a fault of an original computer data storage service gateway node is detected, a new computer data storage service gateway node inquires from the faulty gateway node whether a block device client on the faulty gateway node still opens a data storage cluster volume and mounts a shared directory. If so, an application is made to the data node of the storage cluster to add the faulty gateway node to a blacklist, and the latest blacklist is pushed to all data nodes of the storage cluster, and the blacklist version currently held is periodically reported to the management node of the storage cluster. The new computer data storage service gateway node periodically inquires from the management node of the storage cluster whether the blacklist versions reported by the data nodes of each storage cluster are all the latest. If yes, the computer data storage service is started on the current node, otherwise, the next inquiry cycle is entered after waiting for a period of time.

Description

Technical Field

[0001] The present invention relates to the field of file gateway fault switching, and in particular to a method, device and medium for processing file gateway fault switching. Background Art

[0002] The client communicates with the computer data storage service (NAS Network-attached storage, a file-level (as opposed to block-level storage) computer data storage server that is connected to a computer network and provides data access to heterogeneous network users) gateway node through the service IP to access the computer data storage service. In order to achieve high availability of the computer data storage service, multiple computer data storage service gateway nodes will be configured in the storage system. Each computer data storage service gateway node runs the FDND (FDN Daemon, a daemon process that runs on each node to be detected in FDN) process to form an FDN (Fault detection network, a method for detecting node failures in the network, and supporting the selection of other normal nodes to replace the failed node to provide services) cluster, which is responsible for detecting the operating status of each computer data storage service gateway node. If FDN finds that a computer data storage service gateway node is abnormal, it will select another computer data storage service gateway node in the computer data storage service gateway cluster, assign it a service IP, and let the node take over the failed node to provide computer data storage services.

[0003] The computer data storage service MGR (computer data storage service Manager runs on the computer data storage service gateway node and is responsible for managing the computer data storage service gateway process) process on the computer data storage service gateway node periodically queries the FDN for the status of the current node. If it is found that the status of the current computer data storage service gateway node has changed from normal to abnormal, the computer data storage service on this node will be stopped, and the block device client (KBDC Kernel Block Device Client runs on the kernel's block device client, connects to the storage cluster downward, and provides block device services to the kernel upward) will be notified to unmount the shared directory, and then release access to the data storage cluster volume; if the computer data storage service MGR finds that the current computer data storage service gateway node is selected by FDN to become the new computer data storage service node, it will notify the block device client to establish access to the data storage cluster volume, mount the shared directory, set the service IP and start the NFSD / SMBD (NFS Daemon provides NFS file service daemon process / SMB Daemon provides CIFS file service daemon process) process to start providing computer data storage services. However, in the existing failover scheme, although the block device client volume on the failed computer data storage service gateway node will be unloaded during failover, if the node still has IO blocked in the IO queue, the unloading may fail. In this way, after the network is restored, the failed computer data storage service gateway node will continue to issue IO to the block device client volume, and the block device client volume will be double-written, resulting in data inconsistency. Summary of the invention

[0004] The object of the present invention is to provide a method, device and medium for processing file gateway failover, which can enable a new gateway node and a restored faulty node to dual-write the same data storage cluster volume to ensure data consistency.

[0005] The present invention provides a method for processing file gateway failure switching, comprising the following steps:

[0006] S1. When the failure of the original computer data storage service gateway node is detected, the new computer data storage service gateway node queries the failed gateway node whether the block device client on the failed gateway node still has the data storage cluster volume open. If so, apply to the management node of the storage cluster (MONSRV Monitor Server storage cluster management node, responsible for monitoring and managing data nodes in the cluster) to add the failed gateway node to the blacklist;

[0007] S2. The management node of the storage cluster pushes the latest blacklist to all data nodes of the storage cluster (DATASRV Data Server data nodes of the storage cluster are responsible for managing data and providing data services to clients) according to the application, and the data nodes of the storage cluster periodically report the blacklist version they currently hold to the management node of the storage cluster;

[0008] S3. According to the blacklist of faulty gateway nodes added by the data nodes of the storage cluster, the new computer data storage service gateway node periodically queries the management node of the storage cluster whether the blacklist versions reported by the data nodes of each storage cluster are the latest. If so, the computer data storage service is started on this node. Otherwise, it waits for a period of time and enters the next query cycle until all the data nodes of the storage cluster are updated to the latest blacklist.

[0009] Preferably: step S2 comprises the following steps:

[0010] S2-1. The management node of the storage cluster pushes the latest blacklist to all data nodes of the storage cluster through a broadcast mechanism according to the application;

[0011] S2-2. The data node of the storage cluster periodically reports the blacklist version it currently holds to the management node of the storage cluster based on the received blacklist.

[0012] Preferably: the broadcast mechanism is message communication between different components of the management node of the storage cluster and the data nodes of the storage cluster.

[0013] Preferably: after the faulty gateway node added to the blacklist recovers from the fault, the data nodes of the storage cluster refuse to receive IO (Input / Output read / write request to the storage system) sent by the faulty gateway node, and then the faulty gateway node directly returns the IO to the upper layer application as an error.

[0014] Preferably: the faulty gateway node periodically attempts to stop the computer data storage service on the node, and if successful, queries the management node of the storage cluster whether the current node is in the blacklist, and if so, applies to the management node to remove itself from the blacklist.

[0015] Preferably: the faulty gateway node periodically attempts to stop the computer data storage service on the node, and if successful, queries the management node of the storage cluster whether the current node is in the blacklist, and if so, removes itself from the blacklist, specifically including:

[0016] Detect whether the node is a non-computer data storage service gateway node. If so, detect whether the shared directory is still mounted on the current node. If so, unmount the shared directory and notify the block device client to close the block device client volume.

[0017] Check whether the current node is added to the blacklist, and if so, remove it from the blacklist.

[0018] Preferably: each blacklist version is identified by an incremental version number used by the blacklist.

[0019] The present invention also provides a device for processing file gateway failure switching, the device comprising:

[0020] A computer data storage service gateway node fault detection module is used for, when a fault is detected in the original computer data storage service gateway node, the new computer data storage service gateway node queries the faulty gateway node whether the block device client on the faulty gateway node still opens the data storage cluster volume, and if so, applies to the data node of the storage cluster to add the faulty gateway node to the blacklist;

[0021] A blacklist reporting module, which is used for the management node of the storage cluster to push the latest blacklist to all data nodes of the storage cluster according to the application, and the data nodes of the storage cluster periodically report the blacklist version currently held by them to the management node of the storage cluster;

[0022] The judgment switching module is used to add the faulty gateway node blacklist to the data nodes of the storage cluster. The new computer data storage service gateway node periodically queries the management node of the storage cluster whether the blacklist versions reported by the data nodes of each storage cluster are the latest. If so, the computer data storage service is started on this node. Otherwise, it waits for a period of time and enters the next query cycle until all the data nodes of the storage cluster are updated to the latest blacklist.

[0023] The present invention also provides a non-volatile computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are executed by a server, the server executes the method as described above.

[0024] The present invention provides a method, device and medium for processing file gateway fault switching. During fault switching, a new computer data storage service gateway node first queries whether a block device client on an original faulty node still mounts a shared directory. If the original faulty node still mounts a shared directory, the node is added to a blacklist, thereby ensuring that when the computer data storage service is started on the new computer data storage service gateway node, the original faulty gateway node cannot send a read / write request to a data node of a storage cluster even if the fault is recovered, thereby avoiding double writing of the same data storage cluster volume by the new gateway node and the recovered faulty node, thereby ensuring data consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic diagram of the overall process of fault switching in the prior art;

[0026] Figure 2 A schematic diagram of the behaviors of the original computer data storage service gateway node and the new computer data storage service gateway node during a prior art failover;

[0027] Figure 3 In the process of sending data from NFSD / SMBD to the data storage cluster, each layer may have a cache and queue diagram;

[0028] Figure 4 A schematic diagram of a process for a file gateway failure switch in the prior art;

[0029] Figure 5 A flowchart of a method for processing file gateway failure switching of the present invention;

[0030] Figure 6 A schematic diagram of adding a blacklist to a faulty gateway node of the present invention;

[0031] Figure 7 A schematic diagram of the present invention waiting for data nodes of all storage clusters to obtain the latest blacklist;

[0032] Figure 8 This is a schematic diagram of removing a faulty node from a blacklist according to the present invention;

[0033] Fig. 9 A schematic diagram of a computer data storage service gateway node for initialization and setting of the present invention;

[0034] Fig.10 A schematic diagram of a fault switching when a node of a computer data storage service gateway fails in the present invention;

[0035] Fig.11 This is a schematic diagram of removing a faulty gateway node of the computer data storage service of the present invention from a blacklist. DETAILED DESCRIPTION

[0036] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0037] Before introducing the embodiments of the present invention, the causes of the technical problems in the prior art are first explained with reference to the accompanying drawings. The overall process of the fault switching is as follows: Figure 1 The behavior of the original computer data storage service gateway node and the new computer data storage service gateway node during a failover is as follows. Figure 2 shown.

[0038] When data is sent from NFSD / SMBD to the data storage cluster, it will go through various layers such as Page Cache (Linux file system cache), LFS (local file system), Block device layer (Linux block device abstraction layer) and block device client. Each layer may have cache and queues, such as Figure 3 As shown. If a network error / timeout occurs when sending IO (Input / Output read / write request to the storage system) to the next layer, some kernel components (such as block device client, VFS) will keep retrying or waiting without returning an error to exit; when the communication between the block device client on the computer data storage service gateway node and the data storage cluster is abnormal, the IO will be stuck in a queue in the kernel IO stack. At this time, attempts to close the block device client volume or unmount it will fail. When the communication between the block device client and the data storage cluster is restored, the stuck IO will try to be sent again. If the FDN has detected the gateway node failure and completed the failover at this time, and the client has sent IO from the new gateway node, the IO restored and sent on the failed node may conflict with the IO sent by the client from the new node, resulting in data inconsistency or even damage to the file system structure, such as Figure 4 shown.

[0039] Figure 5The invention is a flow chart of a method for processing a file gateway fault switching provided according to an embodiment of the present invention, comprising S1, when a fault is detected in an original computer data storage service gateway node, a new computer data storage service gateway node queries the faulty gateway node whether a block device client on the faulty gateway node still mounts a shared directory, and if so, an application is made to the data node of the storage cluster to add the faulty gateway node to a blacklist; S2, the management node of the storage cluster pushes the latest blacklist to all the data nodes of the storage cluster according to the application, and the data nodes of the storage cluster periodically report the blacklist version currently held by them to the management node of the storage cluster; S3, according to the blacklist of the faulty gateway node added by the data node of the storage cluster, the new computer data storage service gateway node periodically queries the management node of the storage cluster whether the blacklist versions reported by the data nodes of each storage cluster are the latest, and if so, the computer data storage service is started on this node, otherwise, the next query cycle is entered after waiting for a period of time until all the data nodes of the storage cluster are updated to the latest blacklist. Step S2 includes S2-1, the management node of the storage cluster pushes the latest blacklist to all data nodes of the storage cluster through the broadcast mechanism according to the application; S2-2, the data nodes of the storage cluster periodically report the blacklist version they currently hold to the management node of the storage cluster according to the received blacklist.

[0040] After the original computer data storage service gateway node fails, the new computer data storage service gateway node cannot start the computer data storage service immediately. Instead, it needs to first query whether the block device client on the failed gateway node still mounts the shared directory. If so, it is necessary to add the failed gateway node to the blacklist and wait for all data nodes in the storage cluster to be updated to the latest blacklist before it can start providing computer data storage services. After the failed gateway node network is restored, the block device client attempts to send IO to the data node of the storage cluster and returns a failure. At this time, the block device client will not resend the IO but will directly return the error upward, thereby draining the blocked IO in the IO queue. The subsequent failed gateway node will succeed in attempting to unmount the shared directory and close the block device client volume. Figure 6 shown.

[0041] The blacklist uses an increasing version number to identify each version. The data nodes of the storage cluster in the data storage cluster may have different versions of the blacklist, and they will periodically report the blacklist version numbers they hold to the management node of the storage cluster. After the new computer data storage service gateway node adds the faulty computer data storage service gateway node to the blacklist, it first obtains the latest blacklist version number from the management node of the storage cluster, and then queries the management node of the storage cluster for the blacklist version number held by the data nodes of each storage cluster: If all the data nodes of the storage cluster hold the latest blacklist, the new gateway node can directly start the computer data storage service. If the data nodes of some storage clusters have not yet been pushed the latest blacklist, the new computer data storage service gateway node needs to continue waiting and repeat the query periodically until all the data nodes of the storage cluster have reported the latest blacklist version, at which time the computer data storage service can be started. The waiting process is as follows: Figure 7 shown.

[0042] The computer data storage service MGR process on the faulty gateway node will periodically detect whether the node is a non-computer data storage service gateway node. If so, it will detect whether the block device client volume and mounted shared directory are still open on the current node. If so, it will unmount the shared directory and notify the block device client to close the volume. Then it will detect whether the current node is added to the blacklist. If so, it will apply to the management node to remove itself from the blacklist to ensure that when the subsequent FDN selects the node as the new computer data storage service gateway node, the current node can normally send read and write requests to the data nodes of the storage cluster. The flow chart is as follows: Figure 8 shown.

[0043] Initialize and set up the computer data storage service gateway node

[0044] When a user creates a computer data storage service in the storage system through the storage system management interface, the storage system will set up the computer data storage service gateway node and start the computer data storage service. The process is as follows: Fig. 9 shown.

[0045] 1. When a user creates a computer data storage service on the storage system, the FDN network composed of the FDND processes of the computer data storage service gateway cluster will first select a computer data storage service gateway node as the computer data storage service node ( Fig. 9 The computer data storage service gateway node 1 is selected;

[0046] 2. The computer data storage service MGR on the computer data storage service gateway node 1 periodically queries the FDND process whether the current node is a computer data storage service node. If so, it notifies the block device client to open the data storage cluster volume as a block device client volume, mounts the block device client volume as a shared directory, starts the NFSD / SMBD process, and then sets the service IP to the current node to start providing computer data storage services;

[0047] 3. The computer data storage service MGR on the computer data storage service gateway node 2 will also periodically query the FDND process whether the current node is a computer data storage service node. If it is not, it will wait for a period of time before initiating the next query;

[0048] 4. The request of the computer data storage service client is sent to the computer data storage service gateway node 1 through the service IP. Then the computer data storage service gateway node 1 requests data from the data node of the storage cluster. The data node of the storage cluster queries the blacklist and finds that the gateway node is not on the blacklist. It allows access and returns data to the gateway node. The computer data storage service gateway node 1 then returns the data to the computer data storage service client.

[0049] Failover when a computer data storage service gateway node fails

[0050] When the computer data storage service gateway node that provides the service fails, FDN detects the failure and starts failover. The process is as follows: Fig.10 shown.

[0051] 1. When the gateway node 1 that originally provided computer data storage service fails, the FDN network first detects the failure and then selects gateway node 2 to take over the computer data storage service from gateway node 1;

[0052] 2. On gateway node 1, the computer data storage service MGR periodically queries FDND whether the current node is still the computer data storage service node. After a failover occurs, the computer data storage service MGR learns that the current node is no longer the computer data storage service node, so it releases the service IP on the current node and tries to unmount the block device client volume. However, since there are still unprocessed IO requests remaining in the system when the failure occurs, the unmounting fails; the remaining IO will keep trying to resend to the data nodes of the storage cluster, but will fail.

[0053] 3. On gateway node 2, the computer data storage service MGR periodically queries FDND whether the current node is still the computer data storage service node. After learning that the original computer data storage service node has failed and the current node has been selected as the new computer data storage service node, the gateway node first queries gateway node 1 whether the shared directory on it has been unmounted. After learning that gateway node 1 has not successfully unmounted the shared directory, gateway node 2 applies to the management node of the storage cluster to add gateway node 1 to the blacklist;

[0054] 4. After the management node of the storage cluster adds gateway node 1 to the blacklist, it pushes the new version of the blacklist to all data nodes of the storage cluster. Each data node of the storage cluster periodically reports the blacklist version it currently holds to the management node;

[0055] 5. Gateway node 2 queries the management node of the storage cluster whether all data nodes of the storage cluster hold the latest blacklist version. If not, the query continues until all data nodes of the storage cluster hold the latest blacklist version.

[0056] 6. After confirming that all data nodes of the storage cluster have pushed the latest blacklist, gateway node 2 starts to open the block device client volume on the node and mount the shared directory, start NFSD / SMBD, set the service IP, and start providing computer data storage services;

[0057] 7. At some point in the future, the failure of gateway node 1 is restored, and the remaining IO on it is successfully sent to the data node of the storage cluster. The data node of the storage cluster detects that gateway node 1 has been added to the blacklist, so it rejects its IO request; all rejected IO requests will not be retried but will be returned directly to the upper-level application by gateway node 1. Finally, there is no residual IO on gateway node 1.

[0058] Computer data storage service fault gateway node removed from blacklist

[0059] When the failed computer data storage service gateway node recovers, in order to continue to undertake computer data storage services, it needs to remove itself from the blacklist. The process is as follows: Fig.11 shown.

[0060] 1. After the failed gateway node 1 is restored, the computer data storage service MGR periodically queries whether the current node is a service node. If it is, it unmounts the shared directory on the current node and closes the data storage cluster volume;

[0061] 2. Gateway node 1 then queries the management node of the storage cluster whether the current node is in the blacklist. If so, it applies to the management node of the storage cluster to remove the current node from the blacklist;

[0062] 3. The management node of the storage cluster removes the computer data storage service gateway node 1 from the blacklist and pushes the blacklist to all data nodes of the storage cluster. At this time, the computer data storage service gateway node 1 can be selected by the FDN to provide computer data storage services.

[0063] Special scene processing

[0064] This solution is also applicable to the switching of computer data storage service gateway nodes when the business network is normal.

[0065] The new gateway node will still query the block device client volume opening status of the original service gateway node. If the block device client on the original gateway node still opens the data storage cluster volume and mounts the shared directory, the original gateway node will be added to the blacklist. The process is the same as the service gateway node switching caused by network failure.

[0066] The business network is disconnected and causes a switch

[0067] The network does not recover after a flash outage. The process is the same as switching due to a computer data storage service gateway node failure.

[0068] The network is restored after a flash interruption. The process is the same as switching when the network of the computer data storage service gateway node is normal.

[0069] The new gateway node fails to add the faulty gateway node to the blacklist and needs to keep retrying until it successfully joins the blacklist before the computer data storage service can be started on the new gateway node;

[0070] The new gateway node cannot query the open status of the block device client volume on the original faulty gateway node due to network and other reasons. It needs to keep retrying until the new gateway node can determine the open status of the block device client volume on the faulty gateway node before it can proceed to the next step (adding to the blacklist, starting the computer data storage service, etc.).

[0071] The present invention also provides a device for processing file gateway failure switching, the device comprising:

[0072] The computer data storage service gateway node fault detection module is used for, when detecting the failure of the original computer data storage service gateway node, the new computer data storage service gateway node queries the failed gateway node whether the block device client on the failed gateway node still mounts the shared directory, and if so, applies to the data node of the storage cluster to add the failed gateway node to the blacklist;

[0073] A blacklist reporting module, which is used for the management node of the storage cluster to push the latest blacklist to all data nodes of the storage cluster according to the application, and the data nodes of the storage cluster periodically report the blacklist version currently held by them to the management node of the storage cluster;

[0074] The judgment switching module is used to add the faulty gateway node blacklist to the data nodes of the storage cluster. The new computer data storage service gateway node periodically queries the management node of the storage cluster whether the blacklist versions reported by the data nodes of each storage cluster are the latest. If so, the computer data storage service is started on this node. Otherwise, it waits for a period of time and enters the next query cycle until all the data nodes of the storage cluster are updated to the latest blacklist.

[0075] The present invention also provides a non-volatile computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are executed by a server, the server can execute the method for processing file gateway failure switching in any of the above method embodiments, for example, executing the above described Figure 5 The functions of steps S1 to S3 of the method, and steps S2-1 to S2-2 of the method.

[0076] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims

1. A method for processing file gateway failure switching, characterized in that: The steps include: S1. When the original computer data storage service gateway node is detected to be faulty, the new computer data storage service gateway node queries the faulty gateway node whether the block device client on the faulty gateway node still mounts the shared directory. If so, apply to the management node of the storage cluster to add the faulty gateway node to the blacklist; S2. The management node of the storage cluster pushes the latest blacklist to all data nodes of the storage cluster according to the application, and the data nodes of the storage cluster periodically report the blacklist version they currently hold to the management node of the storage cluster; S3. Based on the faulty gateway node blacklist held by the data nodes of the storage cluster, the new computer data storage service gateway node periodically queries the management node of the storage cluster whether the blacklist versions reported by the data nodes of each storage cluster are the latest. If yes, the computer data storage service is started on this node. Otherwise, it waits for a period of time and enters the next query cycle until all the data nodes of the storage cluster are updated to the latest blacklist. After the faulty gateway node added to the blacklist recovers from the fault, the data nodes of the storage cluster will refuse to receive the IO sent by the faulty gateway node, and then the faulty gateway node will return the rejected IO to the upper-layer application as an error; The faulty gateway node periodically attempts to stop the computer data storage service on the node. If successful, it queries the management node of the storage cluster whether the current node is in the blacklist. If yes, it applies to the management node to remove itself from the blacklist. The faulty gateway node periodically attempts to stop the computer data storage service on the node. If successful, it queries the management node of the storage cluster to see if the current gateway node is on the blacklist. If so, it applies to the management node to remove itself from the blacklist, including: Detect whether the node is a non-computer data storage service gateway node. If so, detect whether the shared directory is still mounted on the current node. If so, unmount the shared directory and notify the block device client to close the underlying data storage cluster volume. Check whether the current node is added to the blacklist. If so, apply to the management node to remove itself from the blacklist.

2. The method for processing file gateway failure switching according to claim 1, characterized in that: The step S2 comprises the following steps: S2-1. The management node of the storage cluster pushes the latest blacklist to all data nodes of the storage cluster through a broadcast mechanism according to the application; S2-2. The data node of the storage cluster periodically reports the blacklist version it currently holds to the management node of the storage cluster based on the received blacklist.

3. The method for processing file gateway failure switching according to claim 2, characterized in that: The broadcast mechanism is a message communication between different components of the management node of the storage cluster and the data nodes of the storage cluster.

4. The method for processing file gateway failure switching according to any one of claims 1 to 3, characterized in that: Each blacklist version is identified by an incremental version number used by the blacklist.

5. A device for processing file gateway failure switching, characterized in that: The device comprises: The computer data storage service gateway node fault detection module is used for, when detecting the failure of the original computer data storage service gateway node, the new computer data storage service gateway node queries the failed gateway node whether the block device client on the failed gateway node still mounts the shared directory, and if so, applies to the management node of the storage cluster to add the failed gateway node to the blacklist; A blacklist reporting module, which is used for the management node of the storage cluster to push the latest blacklist to all data nodes of the storage cluster according to the application, and the data nodes of the storage cluster periodically report the blacklist version currently held by them to the management node of the storage cluster; The judgment switching module is used to determine whether the blacklist versions reported by the data nodes of each storage cluster are the latest according to the blacklist of the faulty gateway nodes added by the data nodes of the storage cluster. If yes, the computer data storage service is started on this node. Otherwise, it waits for a period of time and enters the next query cycle until all the data nodes of the storage cluster are updated to the latest blacklist. After the faulty gateway node added to the blacklist recovers from the fault, the data nodes of the storage cluster will refuse to receive the IO sent by the faulty gateway node, and then the faulty gateway node will return the rejected IO to the upper-layer application as an error; The faulty gateway node periodically attempts to stop the computer data storage service on the node. If successful, it queries the management node of the storage cluster whether the current node is in the blacklist. If yes, it applies to the management node to remove itself from the blacklist. The faulty gateway node periodically attempts to stop the computer data storage service on the node. If successful, it queries the management node of the storage cluster to see if the current gateway node is on the blacklist. If so, it applies to the management node to remove itself from the blacklist, including: Detect whether the node is a non-computer data storage service gateway node. If so, detect whether the shared directory is still mounted on the current node. If so, unmount the shared directory and notify the block device client to close the underlying data storage cluster volume. Check whether the current node is added to the blacklist. If so, apply to the management node to remove itself from the blacklist.

6. A non-volatile computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by the server, the server executes the method according to any one of claims 1 to 4.

Images