CN114063755A

CN114063755A - Power management method, apparatus, control server and medium for storage system

Info

Publication number: CN114063755A
Application number: CN202010756967.8A
Authority: CN
Inventors: 张曙; 吕和栋; 项君广; 阮军
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2020-07-31
Filing date: 2020-07-31
Publication date: 2022-02-18

Abstract

The present disclosure provides a power management method, apparatus, control server and medium for a storage system. The method comprises the following steps: determining M hard disk groups in a power-up state from among a plurality of hard disk groups of the storage system; monitoring the storage space occupancy rate of the M hard disk groups; determining N hard disk groups from the plurality of hard disk groups if the storage space occupancy rate meets a predetermined condition, wherein the M hard disk groups are different from the N hard disk groups; and supplying power to the N hard disk groups. The embodiment of the disclosure further reduces the energy consumption in the cloud storage platform, prolongs the service life of the hard disk, and reduces the resource waste.

Description

Power management method, apparatus, control server and medium for storage system

Technical Field

The present invention relates to the field of cloud computing, and in particular, to a power management method and apparatus, a control server, and a medium for a storage system.

Background

At present, in order to reduce the Total Cost of Ownership (TCO) of the hard disk of the server on the cloud storage platform, the hard disk density as high as possible is generally pursued. Existing servers, once powered on, mean that all hard disks are powered on at the same time. In practice, however, the storage space occupancy of a mechanical hard disk (HDD) is gradually increasing during the life cycle. I.e. very little data is initially stored and therefore the access frequency is very small. Over time, more and more data is stored. When little data is initially stored and access is infrequent, all hard disks will also be powered. The HDD hard disk is characterized in that power consumption is not greatly different between an idle state and a load state. When the HDD hard disk stores little data and much data, power consumption is almost the same. Thus, when the data stored in the HDD hard disk is too little and the access frequency is very small, the power consumption is wasted. Meanwhile, the HDD hard disk has the characteristics of worst energy consumption efficiency and highest failure rate, and is continuously supplied with power, so that failures are more easily caused. In addition, maintaining high-density storage hardware places severe demands on structural load bearing, room cooling capacity, and the like of an Internet Data Center (IDC) room carrying storage servers. If the storage space of the HDD hard disk is left unused, the energy consumption and the resource waste of the IDC machine room are also caused.

In the prior art, a method for overcoming resource waste caused by the idle storage space of an HDD hard disk is to enable the idle HDD hard disk to be in a dormant or idle state. However, the HDD hard disk still consumes power in a sleep or idle state, and the life of the HDD hard disk still suffers.

Disclosure of Invention

In view of this, the embodiments of the present invention are directed to further reduce energy consumption in a cloud storage platform, improve the service life of a hard disk, and reduce resource waste.

According to an aspect of the present disclosure, there is provided a power management method of a storage system, including:

determining M hard disk groups in a power-up state from among a plurality of hard disk groups of the storage system;

monitoring the storage space occupancy rate of the M hard disk groups;

determining N hard disk groups from the plurality of hard disk groups if the storage space occupancy rate meets a predetermined condition, wherein the M hard disk groups are different from the N hard disk groups;

and supplying power to the N hard disk groups.

Optionally, the predetermined condition comprises: the storage space occupancy is greater than a predetermined occupancy threshold.

Optionally, the number of hard disks in at least one hard disk group in the storage system is preset by:

predicting the access frequency of the current prediction period of the storage system according to the historical access record of the storage system before the current prediction period;

and setting the number of hard disks in the hard disk group according to the access frequency.

Optionally, the predicting the access frequency of the current prediction period of the storage system according to the historical access record of the storage system before the current prediction period includes:

according to historical access records of the storage system in a plurality of historical prediction periods before the current prediction period, obtaining the access frequency of the storage system in the plurality of historical prediction periods respectively;

determining the trend of the access frequency changing along with time according to the access frequency of the storage system in the plurality of historical prediction periods;

and acquiring the access frequency of the current prediction period of the storage system according to the change trend.

Optionally, before determining M hard disk groups in a power supply state from among the plurality of hard disk groups of the storage system, the method further includes:

segmenting the storage system into a plurality of storage clusters, and segmenting at least one of the plurality of storage clusters into a plurality of hard disk groups;

designating at least one first group of disks in at least one of the plurality of storage clusters;

receiving data;

forming a plurality of data backups for the data;

a corresponding storage cluster is designated for data backup, which is stored in a hard disk group of the storage cluster.

Optionally, the method further comprises:

monitoring the access frequency of data in the storage system;

moving the plurality of data backups of the data to a first disk group in a respective storage cluster if the access frequency is below a predetermined access frequency threshold.

Optionally, after moving the plurality of data backups of the data to the first hard disk group in the corresponding storage cluster, the method further comprises: and under the condition that the storage data occupancy rate of the first hard disk group is greater than the preset storage data occupancy rate, stopping power supply to the first hard disk group to which at least one data backup in the plurality of data backups of the data is moved.

Optionally, after the power supply to the first hard disk group into which at least one of the plurality of data backups of the data is moved is stopped, the method further includes:

monitoring a frequency of updates to remaining ones of the plurality of data backups of the data;

and if the updating frequency exceeds a preset updating frequency threshold value, restoring power supply to the first hard disk group which stops supplying power.

According to an aspect of the present disclosure, there is provided a power management apparatus of a storage system, including:

a first determination unit configured to determine M disk groups in a power-on state among a plurality of disk groups of the storage system;

the storage space occupancy rate monitoring unit is used for monitoring the storage space occupancy rates of the M hard disk groups;

a second determining unit, configured to determine N hard disk groups from the plurality of hard disk groups if the storage space occupancy satisfies a predetermined condition, where the M hard disk groups are different from the N hard disk groups;

and the power supply control unit is used for supplying power to the N hard disk groups.

Optionally, the storage system includes a plurality of storage clusters, at least one of the plurality of storage clusters includes a plurality of hard disk groups, at least one first hard disk group is designated in at least one of the plurality of storage clusters, and the data stored in the storage system has a plurality of data backups respectively stored in one hard disk group in one storage cluster.

Optionally, the apparatus further comprises:

the access frequency monitoring unit is used for monitoring the access frequency of the data in the storage system;

a data backup transfer unit for transferring the plurality of data backups of the data into a first hard disk group in a corresponding storage cluster if the access frequency is below a predetermined access frequency threshold.

Optionally, the power supply control unit is further configured to stop supplying power to the first hard disk group to which at least one data backup of the plurality of data backups of the data is moved, if the storage data occupancy rate of the first hard disk group is greater than a predetermined storage data occupancy rate.

Optionally, the apparatus further comprises: an update frequency monitoring unit for monitoring an update frequency of remaining data backups of the plurality of data backups of the data; and the power supply control unit is also used for restoring power supply to the first hard disk group which stops supplying power if the updating frequency exceeds a preset updating frequency threshold value.

According to an aspect of the present disclosure, there is provided a storage system control server including:

a memory for storing computer executable code;

a processor for executing the computer executable code to implement the power management method in the storage system as described above.

According to an aspect of the present disclosure, there is provided a computer-readable medium comprising computer-executable code which, when executed by a processor, implements a power management method in a storage system as described above.

In the embodiment of the disclosure, a storage system is divided into hard disk groups, and according to the storage space occupancy rates of the M hard disk groups already in the power supply state, when the storage space occupancy rates meet a predetermined condition, the number N of the hard disk groups which are supplied with power is determined. The initially powered hard disk groups store only a small amount of data. As the data stored increases, the storage space occupancy increases, and when a predetermined condition is met, the powered hard disk groups may be increased to distribute the storage. Therefore, the number of the hard disk groups in the power supply state is continuously increased along with the increase of the stored data amount, the situation that the number of the hard disks for power supply is large when the stored data amount is small and the power consumption of the hard disks is not obviously different when the stored data amount is small and the stored data amount is large is avoided, and unnecessary power consumption, reduction of the service life caused by long-term electrification of the hard disks, and corresponding resource waste of structural bearing of a machine room, refrigerating capacity of the machine room and the like are avoided. Meanwhile, when the number N of the hard disk groups for supplying power is increased, the hard disk groups are used as the granularity, not the whole server, and the number of the hard disks included in the hard disk groups can be adjusted, so that the flexibility of hard disk power supply control is increased. The reason for using the hard disk group as the granularity is to avoid the problem of low efficiency caused by adding one hard disk power supply.

Drawings

The foregoing and other objects, features, and advantages of the disclosure will be apparent from the following description of embodiments of the disclosure, which refers to the accompanying drawings in which:

fig. 1 illustrates an architecture of a cloud storage platform to which the power management method of the disclosed embodiment is applied;

FIG. 2A illustrates a prior art variation of idle hard disk resources in a hard disk during a lifecycle of the hard disk;

FIG. 2B illustrates a change in idle hard disk resources in a hard disk during a lifecycle of the hard disk in an embodiment of the present disclosure;

FIG. 3A illustrates a storage resource organizational chart of a storage system in the prior art;

FIG. 3B illustrates a storage resource organizational diagram of a storage system in an embodiment of the disclosure;

FIG. 4 illustrates a power management control path structure of a server according to one embodiment of the present disclosure;

FIG. 5 illustrates a flow diagram of a method of power management of a storage system according to one embodiment of the present disclosure;

FIG. 6 shows a block diagram of a power management apparatus of a storage system according to one embodiment of the present disclosure;

FIG. 7 illustrates a block diagram of a storage system control server according to one embodiment of the present disclosure.

Detailed Description

The present disclosure is described below based on examples, but the present disclosure is not limited to only these examples. In the following detailed description of the present disclosure, some specific details are set forth in detail. It will be apparent to those skilled in the art that the present disclosure may be practiced without these specific details. Well-known methods, procedures, and procedures have not been described in detail so as not to obscure the present disclosure. The figures are not necessarily drawn to scale.

The cloud storage platform is a platform for providing storage services in a cloud form, and generally based on a distributed file system, high availability of data is realized through multiple copies of dispersed backup. A network structure of a conventional cloud storage platform is shown in fig. 1. The network fabric includes servers 140, access switches 130, aggregation switches 120, and core switches 110.

The servers 140 are processing and storage entities of the cloud storage platform, and the processing and storage of a large amount of data in the cloud storage platform are all completed by the servers 140.

The access switch 130 is a device used to access the server 140 into the cloud storage platform. One access switch 130 accesses multiple servers 140. The access switches 130 are typically located on Top of the Rack, so they are also called set-Top (Top of Rack) switches, which physically connect the servers.

Each aggregation switch 120 connects multiple access switches 130 while providing other services such as firewalls, intrusion detection, network analysis, and the like.

Core switch 110 provides high-speed forwarding for packets entering and exiting the cloud storage platform and connectivity for aggregation switch 120. The network of the entire cloud storage platform is divided into an L3 layer routing network and an L2 layer routing network, and the core switch 110 provides an elastic L3 layer routing network for the network of the entire data center.

Typically, the aggregation switch 120 is the demarcation point between L2 and L3 layer routing networks, with L2 below and L3 above the aggregation switch 120. Each group Of aggregation switches manages a Point Of Delivery (POD), within each Of which is a separate VLAN network. Server migration within a POD does not have to modify the IP address and default gateway because one POD corresponds to one L2 broadcast domain.

FIG. 2A illustrates a prior art variation of idle hard disk resources in a hard disk during the lifecycle of the hard disk. The life cycle refers to the whole process from the putting into use to the damage of the hard disk. The storage space occupancy rate is the ratio of the size of the storage space occupied by the storage data in the hard disk to the total storage space of the hard disk. The graph of fig. 2A is a graph of the storage space occupancy rate as a function of the life cycle when the hard disk is actually used. Note that the curve is not a curve of one hard disk, but an overall curve of all hard disks on the cloud storage platform. In fig. 2A, the hard disk initially stores very little data, and at this time, the occupancy rate of the storage space is very low, so that there are many idle resources, and the data access frequency is very low. The stored data is more and more along with the time, and at the moment, the occupancy rate of the storage space is higher and higher, and the idle resources are reduced. When the hard disk initially stores little data and has a low access frequency, all the hard disks are also powered, and the HDD hard disk is characterized in that the power consumption is not greatly different between the idle state and the load state. Therefore, too low memory occupancy and data access frequency may cause power consumption waste.

In order to overcome the situation that the power consumption is wasted due to the low storage space occupancy rate and the data access frequency, as shown in fig. 2B, in the embodiment of the present disclosure, not all the hard disks need to be powered, and the power supply of each hard disk group is controlled to maintain the access frequency of the data in each hard disk group. At the initial stage of the life cycle, the occupancy rate of the storage space is small, the data access frequency is low, and only a small amount of hard disks are supplied with power, so that the subsequent data can be intensively stored in the small amount of hard disks, and the occupancy rate of the storage space of the hard disk group and the data access frequency are improved. As the storage space occupancy rate begins to increase, the data access frequency increases, and more hard disks are powered to disperse the storage pressure of the powered hard disks. A hard disk that is not powered up has no power consumption. In fig. 2B, the portion sandwiched by the storage space occupancy rate curve and the staircase curve is the idle hard disk resource. It can be seen that, in the embodiment of the present disclosure, idle hard disk resources are greatly reduced, and power consumption waste is reduced.

As shown in FIG. 3A, a prior art storage system 160 includes a plurality of storage clusters 150 and a storage system control server 160. Storage system 160 refers to the portion of replicated storage in a cloud storage platform. As shown in fig. 1, the access switch 130, the aggregation switch 120, and the core switch 110 mainly perform an interaction function, and the server 140 performs a storage function, and therefore, the storage system 160 may be considered to include each server 140 and its control device, i.e., the storage system control server 160. The storage system control server 160 performs control of each server 140. The power management in the embodiment of the present disclosure is mainly performed in the storage system control server 160.

The storage cluster 150 is a cluster of a plurality of servers 160, which corresponds to a cabinet. Each enclosure houses a plurality of servers 160 forming a cluster, and accordingly, an access switch 130 is typically located at the top of the enclosure and is responsible for the exchange of data from the servers 160 to the higher level switches within the cluster. At least one server includes a plurality of hard disks 180 for storing data. To prevent data loss and increase data availability, multiple data backups are typically generated for data to be stored in the storage system 160, as shown in FIG. 3A, each stored to one hard disk 180 of one server 140 in multiple storage clusters 150. Once one of the data backups is lost, the other data backups may be found in the hard disks 180 of the other storage clusters 150.

FIG. 3B illustrates a storage resource organizational chart of the storage system in an embodiment of the disclosure. As compared to fig. 3A, it introduces the concept of hard disk groups. The hard disk group is a set of a plurality of hard disks and is the minimum unit for increasing and stopping power supply. The plurality of hard disks in a hard disk group are simultaneously powered on or powered off. One storage cluster 150 includes a plurality of servers 140, one server 140 including a plurality of hard disk groups 170, one hard disk group including a plurality of hard disks 180.

According to one embodiment of the present disclosure, a power management method of a storage system is provided. It is performed by the storage system control server 160 of fig. 3B. Power management refers to controlling the starting and stopping of power supply to a hard disk in a storage system. As shown in fig. 5, the method includes:

step 410, determining M hard disk groups in a power supply state in a plurality of hard disk groups of the storage system;

step 420, monitoring the occupancy rate of the storage space of the M hard disk groups;

step 430, if the storage space occupancy rate meets a predetermined condition, determining N hard disk groups from the plurality of hard disk groups, wherein the M hard disk groups are different from the N hard disk groups;

and step 440, supplying power to the N hard disk groups.

In a storage system, only m hard disk groups 170 may be initially powered up in each storage cluster 150. In this case, M is M. Then, the storage space occupancy rate of the hard disk group is detected, and once a predetermined condition is met, power supply is added to N hard disk groups, namely N is equal to N. Since the power management method of the disclosed embodiment is performed periodically. In the next period, monitoring that M is M + N, once the occupancy rate of the storage space of the M + N hard disk groups meets a preset condition, increasing N hard disk groups for supplying power, namely N is N, and so on.

In the above scheme, the reason why M hard disk sets 170 are found in each storage cluster 150 to supply power is not that some storage clusters 150 have hard disk sets 170 to supply power, and some storage clusters 150 do not have data backup because data exists in different storage clusters 150, so that the number of hard disk sets 170 supplying power in each storage cluster 150 is required to be balanced, and excessive hard disk sets supplying power in one storage cluster 150 and too little hard disk sets supplying power in another storage cluster 150 are avoided as much as possible, which is beneficial to load balancing.

At this point, since one or more of the hard disk groups 170 in each storage cluster 150 are powered, the backup of data stored to the storage cluster 150 is stored to the one or more powered hard disk groups 170. Then, in step 410, the M hard disk groups in the powered state are determined. The storage system control server 160 has a record of powering up the disk group 170, and thus, it is possible to determine, by looking up the record, M disk groups in a powered-up state among the plurality of disk groups of the storage system.

In step 420, the storage occupancy of the hard disk group 170 in the powered state is monitored. If the preset condition is not met, the stored data does not reach a certain level, and power supply of other hard disk groups is added, so that power consumption is wasted. Therefore, in step 430, the new N hard disk groups are powered up only when the stored data reaches a certain level (predetermined conditions are met). In this way, the backup of data stored to the storage cluster 150 is later dispersed across the original hard disk group 170 and the additional N hard disk groups 170. The total storage occupancy of the powered hard disk clusters 170 is then monitored after a period of use.

In step 430, if the storage occupancy satisfies a predetermined condition, N hard disk groups are determined from the plurality of hard disk groups, wherein the M hard disk groups are different from the N hard disk groups. That is, the N disk groups are determined from among the disk groups not in a powered state from among the plurality of disk groups.

In step 440, power is supplied to the N hard disk groups. Steps 410-440 are performed periodically. In the execution of step 440, in the next cycle, step 410 is executed again to determine the hard disk group in the power supply state, and so on, and the loop is executed.

The storage space occupancy rate is a ratio of a size of a storage space occupied by storage data in the hard disk in the power supply state to a size of a total storage space of the hard disk in the power supply state. If one hard disk group is in the power supply state, the hard disks in the power supply state are all the hard disks in the hard disk group. If a plurality of hard disk groups are in the power supply state, the hard disks in the power supply state are all the hard disks in the plurality of hard disk groups.

In one embodiment, step 420 may be performed by the storage system control server 160 periodically polling the server 140 for a hard disk group that is in a powered-on state. The storage system control server 160 periodically initiates a query request to each server 140. After receiving the query request, the server 140 returns the storage space occupancy rate of the hard disk group in the power supply state to the storage system control server 160.

In one embodiment, the predetermined condition may be that the storage occupancy is greater than a predetermined occupancy threshold. If the occupancy rate of the storage space is greater than the preset occupancy rate threshold, the ratio of the occupancy rate of the storage data in the hard disk in the power supply state to the total capacity of the hard disk in the power supply state is more than the threshold, and a new hard disk group can be allocated for sharing. If not shared, although the power consumption is sufficiently small, it is easy to cause overflow. Thus, the number of hard disk groups to be powered can be increased.

In one embodiment, the number of hard disks in at least one group of hard disks in the storage system may be set in advance by: predicting the access frequency of the current prediction period of the storage system according to the historical access record of the storage system before the current prediction period; and setting the number of hard disks in the hard disk group according to the access frequency.

In the disclosed embodiment, the time line of prediction is divided into time intervals, and prediction is performed at regular time intervals, which are called prediction periods. For example, at a fixed time interval of a month, the current prediction period is the interval from the current time point to a time point one month after the current time point. The current prediction period is preceded by the current point in time. The access frequency refers to the total access times of the user to access the storage system in a prediction period.

In one embodiment, predicting the access frequency of the current prediction period of the storage system according to the historical access records of the storage system before the current prediction period may include the following processes:

acquiring the access frequency of the storage server group in a plurality of historical prediction periods according to historical access records of the storage server group in the plurality of historical prediction periods before the current prediction period;

determining the trend of the access frequency changing along with time according to the access frequency of the storage server group in the plurality of historical prediction periods;

and acquiring the access frequency of the current prediction period of the storage server group according to the change trend.

The historical prediction period refers to a historical prediction period before the current time point, namely, a period formed by pushing forward the current time point by 1, 2, 3 and 4 … … prediction period lengths. For example, at regular time intervals of months, a historical prediction period … … is formed between a time point one month before the current time point and the current time point, and a historical prediction period … … is formed between a time point two months before the current time point and a time point one month before the current time point

The historical access records in a plurality of historical prediction periods before the current prediction period comprise the time of each access of the user to the hard disk, the visitor, the accessed hard disk and other information related to the access in the plurality of historical prediction periods. And counting the number of all the historical access records to obtain the access frequency of the storage server group in the plurality of historical prediction periods. At this time, the time variation trend of the access frequency can be determined according to the access frequency of the storage server group in the plurality of historical prediction periods. This trend may be embodied as a time-dependent profile of the access frequency. Points are drawn in a coordinate system by taking the time point as a horizontal axis and the access frequency as a vertical axis, and the points are connected to obtain an access frequency variation curve with time, wherein the variation curve shows the variation trend of the access frequency with time. Since only the condition of a plurality of historical prediction periods before the current prediction period is drawn, information after the current time point does not exist, but the access frequency change condition after the current time point is predicted by making a tangent line on the curve to extend to the right at a point corresponding to the current time point on the curve. On the curve extending in the tangent line, the abscissa is the point of a prediction period after the current time point, and the ordinate of the point is the access frequency of the current prediction period of the storage server group.

By predicting the access frequency of the current prediction period of the storage system, the number of the hard disks in the hard disk group can be set according to the access frequency.

In one embodiment, the number of hard disks is proportional to the access frequency. That is, the greater the access frequency, the greater the number of hard disks in each hard disk group, i.e., the greater the granularity. Since the larger the access frequency, the more hard disks that need to be added for it per unit time, otherwise it is difficult to cope with such a high access frequency. The effect of this is to make the granularity of hard disk increase each time related to the number of times of user unit time access, reduce the situation that can't meet the user access requirement.

In the above embodiment, in the case that the storage system stores not much data, the number of hard disk sets to be powered on is small, so that too much energy consumption is not wasted. In another embodiment, it is considered that in some cases, although the storage system stores a relatively large amount of data, the data are all data that are not accessed frequently, and each data on the cloud storage platform has multiple data backups stored in different storage clusters, which also causes waste of energy consumption. Therefore, in this embodiment, data is aggregated in terms of data access frequency, and power supply to the hard disk group in which part of data is backed up is stopped for some data with lower access frequency. That is, since the access frequency thereof is low, it is not necessary to maintain an accessible state for every backup thereof at all times, thereby reducing power consumption.

In one embodiment, prior to step 410, the method may further comprise: segmenting the storage system into a plurality of storage clusters 150, segmenting at least one storage cluster 150 of the plurality of storage clusters 150 into a plurality of hard disk groups 170; designating at least one first group of disks in at least one of the storage clusters 150; receiving data; forming a plurality of data backups for the data; the corresponding storage cluster 150 is designated for data backup, which is stored in a hard disk group 170 of the storage cluster 150.

As shown in FIG. 3B, the storage system is partitioned into a plurality of storage clusters 150, and at least one storage cluster 150 of the plurality of storage clusters 150 is partitioned into a plurality of hard disk groups 170. At least one first group of disks is designated in at least one of the storage clusters 150. The first group of disks is a group of disks dedicated to storing data that is not accessed frequently, e.g., a group of disks storing data having an access frequency below a predetermined access frequency threshold. These data are accessed infrequently and do not necessarily take up too many resources. These data are placed into the first hard disk group. A part of these first hard disk packs may be powered down in a later process to reduce power consumption.

When the storage system receives the data, a plurality of data backups (e.g., 3 data backups) are formed for the data. Each data backup is assigned a storage cluster 150, and the data backup is stored in a hard disk group 170 of the storage cluster 150. As shown in FIG. 3B, storage cluster 1 is selected for data backup 1 and stored in one of the hard disk groups 170 in storage cluster 1; selecting storage cluster 2 for data backup 2, storing it in one of the hard disk groups 170 in storage cluster 2; selecting storage cluster n for data backup 3, storing it in one of the hard disk groups 170 in storage cluster n; the storage clusters 3 to n-1 do not store any backup of this data.

Since there is a first disk group in each storage cluster 150, if a certain data is accessed infrequently, its respective backup may be transferred to the respective first disk group. A portion of these first hard disk stacks may be selected to be de-energized to reduce power consumption. For example, there are 3 data backups of a certain infrequently accessed data, each stored in the first disk group of each of the 3 storage clusters. Because the data backup system is accessed infrequently, the first hard disk group where 2 data backups in 3 data backups are located can stop supplying power, and energy consumption is reduced. Because 1 data backup is not powered off, other data backups can be recovered at any time.

In this embodiment, the method further comprises:

monitoring the access frequency of data in the storage system;

In one embodiment, monitoring the frequency of access to data in the storage system may be by way of the storage system control server 160 periodically polling the server 140. The storage system control server 160 periodically initiates periodic query requests to the servers 140 in the storage system. After receiving the query request, each server 140 obtains the number of times that each data stored therein is accessed within a predetermined time period, i.e., the access frequency, from the access log according to the access log stored therein, and sends the number of times to the storage system control server 160. For example, from the access log, it can be obtained how many times each data is accessed in the time period between the time point 24 hours before the current time point and the current time point, that is, the access frequency of each data.

Then, for each data, it is determined whether its access frequency is below a predetermined access frequency threshold. If so, it is indicated as cold data, which is infrequently accessed and may be included in the first hard disk group. Since the data has multiple data backups, each data backup is stored in a different storage cluster 150, as long as the data backup in one storage cluster 150 is assigned to the first hard disk group of the storage cluster 150, the rest of the data backups will also be assigned to the first hard disk group of the corresponding storage cluster 150.

In one embodiment, the method further comprises: and under the condition that the storage data occupancy rate of the first hard disk group is greater than the preset storage data occupancy rate, stopping power supply to the first hard disk group to which at least one data backup in the plurality of data backups of the data is moved. This is because if the storage data occupancy of the first hard disk group is small, it is not necessary to power it down. The storage data occupancy rate is equal to the ratio of the storage space of the stored data in the first hard disk group to the whole storage space of the first hard disk group. When the ratio is larger than the preset storage data occupancy rate, the storage data is enough, the frequency of the data is low, and the first hard disk group in which a part of data is backed up can be considered to be powered off. If the access frequency of the data is not high, a plurality of data backups are set for the data to be stored respectively, and one storage space is occupied, so that the energy consumption is increased. Therefore, the first hard disk group occupied by a portion of the plurality of data backups in the respective storage cluster 150 may be powered down. For example, in the case of saving 3 data backups, the first hard disk set occupied by 1-2 data backups may be powered off.

In one embodiment, after the power supply to the first hard disk group into which at least one of the plurality of data backups of the data is moved is stopped, the method further comprises:

The remaining copies, i.e., the data copies of the plurality of data copies excluding the portion of the data copies on the first hard disk group to which power was stopped, i.e., the data copies on the hard disk groups to which power was not stopped. For example, the power supply is stopped for the first hard disk set occupied by 2 data backups of 3 data backups of the same unusual data, and the rest data backups refer to the data backups on the hard disk sets which are not stopped. The update frequency refers to the frequency of being overwritten. The access includes a read or a write. The hard disk set is powered off, so that reading is not influenced, and as long as a residual data backup exists, the data can be read subsequently. However, if the access is a write and the first hard disk set in which a portion of the plurality of data backups is located has been powered down, then a situation may occur in which the remaining data backups are overwritten and the portion that has been powered down is not overwritten. The hard disk group which stops supplying power may recover power in the future, so that different data backups of the same data are inconsistent. In order to reduce the occurrence of the situation, the updating frequency of the rest data backups in the plurality of data backups is monitored, the updating frequency refers to the frequency of being overwritten, and once the frequency is higher, the first hard disk group which stops supplying power is immediately restored to supply power.

In one embodiment, monitoring the update frequency may be by way of the storage system control server 160 periodically polling the server 140. The storage system control server 160 periodically initiates a periodic query request to all servers 140 in all storage clusters 150. After receiving the query request, the server 140 obtains the number of times that the server is overwritten in unit time, that is, the access frequency, from the access log according to the access log stored in the server, and sends the access frequency to the storage system control server 160.

The storage system control server 160 then determines whether the update frequency exceeds a predetermined update frequency threshold. And if the updating frequency exceeds a preset updating frequency threshold value, restoring power supply to the first hard disk group which stops supplying power.

In the embodiment of the disclosure, if unnecessary energy consumption is large due to the fact that stored data are mainly low in access frequency, power supply can be stopped for a hard disk set occupied by a part of data backup of some data with low access frequency, and when the update frequency is large enough, power supply can be resumed. Therefore, unnecessary energy consumption and resource consumption caused by low data access frequency in the hard disk group are reduced.

In the above process, monitoring the access frequency of the data and monitoring the update frequency of the remaining data backups in the plurality of data backups may be performed periodically, so that the change condition of the cold and hot data (data with low access frequency and data with high access frequency) in the storage system can be tracked in time, and the condition that the cold data becomes hot data and the condition that the hot data becomes cold data can be found in time.

In the above process, the addition of the disk group to be powered, the stop of the power supply to a part of the first disk group, and the restoration of the power supply when the update frequency exceeds the predetermined update frequency threshold are controlled by a path inside the server 140. The storage system control server 160 may collect the switch states of the respective hard disk groups through a path and control the on and off of the switches of the respective hard disk groups according to the collected switch states and whether power supply is to be performed or stopped. The switch state refers to whether the hard disk group switch is on or off. This path is described in detail below.

As shown in fig. 4, the server 140 is divided into a software layer 1412 and a hardware layer 1413. The hardware layer 1413 is divided into a motherboard 1414 and a memory module 1415.

Software layer 1412 includes distributed file system control agent 1416, operating system 1417, disk drive 1418, and Basic Input Output System (BIOS) 1420. The distributed file system control agent 1416 is application software that is provided in the storage server 140 opposite the storage system control server 160 and that performs the function requested by the storage system control server 160. The operating system 1417 is a computer program that manages the hardware and software resources of the server 140, and is also the kernel and foundation of the server 140. The disk drive 1419 is a component that drives the disk controller 1423 so that the disk operates normally. The basic input/output system (BIOS)1420, which is a standard firmware interface in the industry on IBM PC compatible systems, stores the most important basic input/output programs of the server 140, the post-power-on self-test program, and the system self-start program, and can read and write specific information set by the system from the CMOS.

The motherboard 1414 includes a disk controller 1423, a platform-controlled multi-port transponder (PCH)1421, and a Baseboard Management Controller (BMC) 1422. The disk controller 1423 is a device that controls reading and writing of a disk. The PCH 1421 is a device for preventing a line fault in the server 140, and it can not affect the operation of other lines when a fault occurs in one line in the server 140. The BMC 1422 performs basic control functions for the memory module 1415.

The memory module 1415 includes a Complex Programmable Logic Device (CPLD)1424 and hard disk bank switches 1425. The CPLD 1424 transmits a power supply/stop control signal to each hard disk group switch 1425, and controls on/off of each hard disk group switch 1425, thereby controlling power supply or stopping power supply. In addition, each hard disk group switch 1425 also reports a power supply/power supply stop state acquisition signal to the CPLD 1424, where the power supply/power supply stop state acquisition signal indicates whether the corresponding hard disk group switch is on or off, that is, whether power supply is supplied or stopped. In addition, each hard disk group switch 1425 is also connected to a power supply.

In one embodiment, the collecting of the switch states of the respective hard disk groups is performed through a first path. The first path contains the CPLD 1424, BMC 1422, disk controller 1423, disk drive 1419, operating system 1417, distributed file system control agent 1416. Each hard disk group switch 1425 also reports a power supply/power supply stop state acquisition signal to the CPLD 1424, the CPLD 1424 reports the power supply/power supply stop state acquisition signal to the distributed file system control agent 1416 through the BMC 1422, the disk controller 1423, the disk drive 1419, and the operating system 1417 in sequence, and finally reports the power supply/power supply stop state acquisition signal to the storage system control server 160 through the corresponding distributed file system control agent 1416. In this way, the inherent components of the server 140 are fully utilized to collect the switch states of the hard disk groups, and the purposes of simplicity and high efficiency are achieved.

In one embodiment, the controlling of the on and off of the hard disk group switches is performed through a second path. The second path contains the distributed file system control agent 1416, operating system 1417, disk drive 1419, BMC 1422, CPLD 1424. The storage system control server 160 sends a power supply/stop control signal to the distributed file system control agent 1416, the distributed file system control agent 1416 sends the power supply/stop control signal to the CPLD 1424 sequentially through the operating system 1417, the disk drive 1419, and the BMC 1422, and the CPLD 1424 sends the power supply/stop control signal to each hard disk group switch 1425, so that the purpose of controlling the power supply/stop of each hard disk group is achieved with a simple circuit.

As shown in fig. 6, according to an embodiment of the present disclosure, there is provided a power management apparatus 500 of a storage system, the apparatus 500 including:

a first determining unit 510, configured to determine M hard disk groups in a power-up state among a plurality of hard disk groups of the storage system;

a storage occupancy rate monitoring unit 520, configured to monitor the storage occupancy rates of the M hard disk groups;

a second determining unit 530, configured to determine N hard disk groups from the plurality of hard disk groups if the storage space occupancy satisfies a predetermined condition, where the M hard disk groups are different from the N hard disk groups;

and a power supply control unit 540, configured to supply power to the N hard disk groups.

Optionally, the apparatus 500 further comprises:

Optionally, the power supply control unit 540 is further configured to stop supplying power to the first hard disk group to which at least one data backup of the plurality of data backups of the data is moved, if the storage data occupancy rate of the first hard disk group is greater than the predetermined storage data occupancy rate.

Optionally, the apparatus 500 further comprises: an update frequency monitoring unit for monitoring an update frequency of remaining data backups of the plurality of data backups of the data; the power supply control unit 540 is further configured to resume power supply to the first hard disk group that stops supplying power if the update frequency exceeds a predetermined update frequency threshold.

The implementation details of the power management apparatus 500 are fully described in the foregoing embodiments of the method, and are not repeated for brevity.

A power management method of a storage system according to an embodiment of the present disclosure may be implemented by the storage system control server 160 of fig. 7. The storage system control server 160 according to an embodiment of the present disclosure is described below with reference to fig. 7. The storage system control server 160 shown in fig. 7 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

As shown in fig. 7, storage system control server 160 is in the form of a general purpose computing device. The components of storage system control server 160 may include, but are not limited to: the at least one processing unit 810, the at least one memory unit 820, and a bus 830 that couples the various system components including the memory unit 820 and the processing unit 810.

Wherein the storage unit stores program code that is executable by the processing unit 810 to cause the processing unit 810 to perform the steps of the various exemplary embodiments of the present disclosure described in the description section of the above exemplary methods of the present specification. For example, the processing unit 810 may perform the various steps as shown in fig. 5.

The storage unit 820 may include readable media in the form of volatile memory units such as a random access memory unit (RAM)8201 and/or a cache memory unit 8202, and may further include a read only memory unit (ROM) 8203.

The storage unit 820 may also include a program/utility 8204 having a set (at least one) of program modules 8205, such program modules 8205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.

Bus 830 may be any of several types of bus structures including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

The storage system control server 160 may also communicate with one or more external devices 700 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the storage system control server 160, and/or with any devices (e.g., router, modem, etc.) that enable the storage system control server 160 to communicate with one or more other computing devices. Such communication may occur via input/output (I/O) interfaces 850. Also, the storage system control server 160 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN) and/or a public network, such as the Internet) via the network adapter 860. As shown, the network adapter 860 communicates with the other modules of the storage system control server 160 via a bus 830. It should be appreciated that although not shown in the figures, other hardware and/or software modules may be used in conjunction with storage system control server 160, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

It should be understood that the above-described are only preferred embodiments of the present disclosure, and are not intended to limit the present disclosure, since many variations of the embodiments described herein will occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

It should be understood that the embodiments in this specification are described in a progressive manner, and that the same or similar parts in the various embodiments may be referred to one another, with each embodiment being described with emphasis instead of the other embodiments.

It should be understood that the above description describes particular embodiments of the present specification. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

It should be understood that an element described herein in the singular or shown in the figures only represents that the element is limited in number to one. Furthermore, modules or elements described or illustrated herein as separate may be combined into a single module or element, and modules or elements described or illustrated herein as single may be split into multiple modules or elements.

It is also to be understood that the terms and expressions employed herein are used as terms of description and not of limitation, and that the embodiment or embodiments of the specification are not limited to those terms and expressions. The use of such terms and expressions is not intended to exclude any equivalents of the features shown and described (or portions thereof), and it is recognized that various modifications may be made within the scope of the claims. Other modifications, variations, and alternatives are also possible. Accordingly, the claims should be looked to in order to cover all such equivalents.

Claims

1. A method of power management for a storage system, comprising:

monitoring the storage space occupancy rate of the M hard disk groups;

and supplying power to the N hard disk groups.

2. The method of claim 1, wherein the predetermined condition comprises: the storage space occupancy is greater than a predetermined occupancy threshold.

3. The method of claim 1, wherein the number of hard disks in at least one group of hard disks in the storage system is preset by:

4. The method of claim 3, wherein predicting the access frequency of the current prediction cycle of the storage system based on historical access records of the storage system prior to the current prediction cycle comprises:

5. The method of claim 1, wherein prior to determining the M hard disk groups of the plurality of hard disk groups of the storage system that are in a powered-on state, the method further comprises:

receiving data;

forming a plurality of data backups for the data;

6. The method of claim 5, further comprising:

monitoring the access frequency of data in the storage system;

7. The method of claim 6, wherein after moving the plurality of data backups of the data to a first hard disk group in a respective storage cluster, the method further comprises:

and under the condition that the storage data occupancy rate of the first hard disk group is greater than the preset storage data occupancy rate, stopping power supply to the first hard disk group to which at least one data backup in the plurality of data backups of the data is moved.

8. The method of claim 7, wherein after powering down the first hard disk set into which at least one of the plurality of data backups of the data was moved, the method further comprises:

9. A power management apparatus of a storage system, comprising:

10. The apparatus of claim 9, wherein the predetermined condition comprises: the storage space occupancy is greater than a predetermined occupancy threshold.

11. The apparatus of claim 9, wherein the number of hard disks in at least one group of hard disks in the storage system is preset by:

12. The apparatus of claim 11, wherein predicting the access frequency of the current prediction cycle of the storage system based on historical access records of the storage system prior to the current prediction cycle comprises:

13. The apparatus of claim 9, wherein the storage system comprises a plurality of storage clusters, at least one of the plurality of storage clusters comprising a plurality of hard disk groups, at least one first hard disk group being designated in the at least one of the plurality of storage clusters, the data stored in the storage system having a plurality of data backups each stored in one of the hard disk groups in one of the storage clusters.

14. The apparatus of claim 9, further comprising:

15. The apparatus of claim 14, wherein the power supply control unit is further configured to stop supplying power to the first hard disk group to which at least one of the plurality of data backups of the data is moved if the storage data occupancy of the first hard disk group is greater than a predetermined storage data occupancy.

16. The apparatus of claim 15, further comprising: an update frequency monitoring unit for monitoring an update frequency of remaining data backups of the plurality of data backups of the data;

and the power supply control unit is also used for restoring power supply to the first hard disk group which stops supplying power if the updating frequency exceeds a preset updating frequency threshold value.

17. A storage system control server, comprising:

a memory for storing computer executable code;

a processor for executing the computer executable code to implement the power management method of the storage system according to any one of claims 1 to 8.

18. A computer-readable medium comprising computer-executable code which, when executed by a processor, implements a method of power management for a storage system according to any of claims 1-8.