1323414 九、發明說明： 【發明所屬之技術領域】 一種建立快照之方法，特別是指一種以時間戳分隔連續的儲 存空間來集中建立依賴型快照之方法。 【先前技術】 快疋-種可以對儲存區域中某一特定時間點的資料進行即 時備份，並幾乎不影響資料服務（也就是㈣讀寫）效率的= 方式，匕疋一個獨立存在於儲存媒體上的儲存區域。快照的不β 對儲存區域的資料完整備份，而是將儲存區域分為數個小區^ ，通常以ΚΒ為單位，其大小是可以調整的），當使用者 =寫儲存區域的資料時’把將被更改的區塊的資料複製到快昭 二:=是輪製驗騎於是，當一： 立時’使用者需要指定使_碟空間、對應的儲存區 成專貝訊，快照會備份其建立時間點時的資料。 儲存建錢健照，純__不_時間點的 _ ^域的諸’麵輯和物理上都佔 了避免當對儲存區域建立了數個快照時，各二 相同資料複製到各自的錯存空間中 /、…曰把將被更改的 並加重了儲存媒體進行續 4了錯存空間的浪費， 較早建立的快照依賴^’貞擔’所以各快照有依賴的機制： 者更改的轉只會被複細最新、晚^ ’也就是說，被使用 有效的避免了鱗郎的 ^的贿、巾，這樣就 減I了儲存嬅體讀寫的負擔。 1323414 雖然將儲存區域建立數個快照可以備份各個時間點的資料， 但是，快照卻會造成空間浪費的問題。由於各快照是獨立的儲存 區域，卻又有依賴關係，所以當新的快照被建立時，已存在的快 照將不再存入任何資料，如「第]圖」所示，儲存區域彻建立 ;了第一快照121、第二快照122、第三快照123等三個快照，第 ；一快照121僅有第一已使用區域121a儲存資料’第二快照122 禮以第二已使用區域122a錯存資料，第三快照η3亦同，由於 •第三未使用區域123b在新的快照被建立之前還會有資料存入， 是故，第-未使用區域121b與第二未使用區域122b形成了儲存 媒體100中的被浪費的儲存空間。因此，如何能提供一種可減少 儲存空間的浪費的快照，成為待解決的問題。 【發明内容】 馨於以上_題，本發明的主要目的在於提供—種可以將儲 存空間配置給對應儲存區域的數個依賴独照，使得各快照使用 釀I塊賴賴存^ ’達職依_快照沒絲儲存㈣區塊的 空間，藉以解決先前技術所提到之儲存空間被浪費之問題。 為達上述目的，本發明所揭露之方法，包括有下列步驟：建 儲存ϋ域之第糾?、、儲存具有第—快照的起始位址與第一快 .‘、、、的建立時間之第-時間戮，並奴寫人前複製之—指標為第一 、…、的起始位址、§儲存區域中之資料區塊（比^他）將被改變時’ 執仃寫複製’並於寫人前複製完成後，修改寫人前複製之指 f為該第-快照中可繼續儲存龍區塊之位址、以修改後的寫入 前複製之指標為儲存區域之第二快照的起始位址，建立第二快 1323414 照’並儲存第二快照的起始位址與第二快照的建立時間為第二時 間戳。 有關本發明之詳細特徵與實作，茲配合圖示在實施方式中詳 細說明如下’其内容足以使任何熟習相關技藝者了解本發明之技 術内容並據以實施’且根據本說明書所揭露之内容及圖式，任何 熟習相關技蟄者可輕易地理解本發明相關之目的及優點。 【實施方式】1323414 IX. Description of the invention: [Technical field to which the invention pertains] A method of establishing a snapshot, in particular, a method of collectively establishing a dependent snapshot by separating successive storage spaces by time stamps. [Prior Art] Quick-time data can be instantly backed up at a specific point in time in the storage area, and it hardly affects the efficiency of the data service (that is, (4) reading and writing) = 匕疋 an independent existence in the storage medium Storage area on. The snapshot is not a complete backup of the data in the storage area, but the storage area is divided into several cells ^, usually in units of ,, the size can be adjusted), when the user = write the data of the storage area The data of the changed block is copied to the fast Zhao 2: = is the round test, then when: one immediately: the user needs to specify the _ disc space, the corresponding storage area into a special Beixun, the snapshot will back up its settling time point Time information. Storing Jian Jian Jian Zhao, pure __ not _ time point _ ^ domain of the 'face and physical account for the occupation to avoid the creation of several snapshots of the storage area, the two identical data copied to their respective In the space /, ... 曰 will be changed and aggravated the storage media to continue the 4 waste of space, the earlier established snapshot depends on ^ '贞 ' 所以 所以 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各 各It will be the most recent, late ^ 'that is, it is used to effectively avoid the bribes and towels of the scales, thus reducing the burden of storing the body reading and writing. 1323414 Although it is possible to back up data at various points in time by creating several snapshots of the storage area, snapshots cause a waste of space. Since each snapshot is a separate storage area, but has dependencies, when a new snapshot is created, the existing snapshot will no longer store any data, as shown in the "Figure", the storage area is completely established; The first snapshot 121, the second snapshot 122, and the third snapshot 123 have three snapshots. The first snapshot 121 stores only the first used area 121a. The second snapshot 122 is saved in the second used area 122a. The data, the third snapshot η3 is also the same, since the third unused area 123b has data stored before the new snapshot is created, the first unused area 121b and the second unused area 122b are stored. Wasted storage space in the media 100. Therefore, how to provide a snapshot that can reduce the waste of storage space becomes a problem to be solved. SUMMARY OF THE INVENTION The main purpose of the present invention is to provide a plurality of dependent photos that can allocate a storage space to a corresponding storage area, so that each snapshot uses a block of I. _Snapshots do not store the space in the (4) block to solve the problem of wasted storage space mentioned in the prior art. In order to achieve the above objective, the method disclosed in the present invention includes the following steps: constructing a storage area, and storing a start address of the first snapshot and a setup time of the first fast. ', , , The first-time 戮, and the slave copy is copied before - the index is the starting address of the first, ..., § the data block in the storage area (ratio ^ he) will be changed when 'transfer copy' and After the copying is completed, the finger f before the copy is modified is the address of the first block in which the dragon block can continue to be stored, and the modified copy before writing is the start position of the second snapshot of the storage area. Address, establish a second fast 1323414 and 'save the start address of the second snapshot and the set time of the second snapshot as the second timestamp. Detailed Description of the Invention The detailed description and implementation of the present invention are described in detail in the following description, which is set forth in the accompanying claims And the related art and the related objects and advantages of the present invention can be easily understood by those skilled in the art. [Embodiment]
以下以一個實施例與來解說本發明的運作系統與方法，本實 施例將以具有邏輯卷管理（L〇gica丨Volume Management; LVIVI) 技術之儲存系統為例，但並不以此為限，其他儲存系統例如磁碟 陣列（RedundantArrayof 丨nexpensiveDisk;RA|D)亦可使用執 行有本發明之㈣程式。LVM可靖所有的儲存設備統一管理， 意即LVM會將所有的儲存媒體整合為一個大型的儲存空間，稱之The operating system and method of the present invention will be described below with reference to an embodiment. The storage system of the technology of Logical Volume Management (LVIVI) is taken as an example, but not limited thereto. Other storage systems such as a disk array (Redundant Array of expnexpensive Disk; RA|D) may also be used to execute the program of the present invention. LVM can manage all storage devices in a unified manner, meaning that LVM will consolidate all storage media into one large storage space, called
為卷組（Volume Group; VG)，使用者可以在卷組中指定特定偉 存空間的儲存區域’ LVM會將其劃分為—個邏輯卷（L〇gjca Volume; LV)，卷組中有一個特定的儲存區域错存著各個邏 分酉嘛组中的位址，如「第1圖」所示，儲存空間彻即為卷 組’第-儲存區域110、第一快照121、第二快照122、第三快照 123為邏輯卷’而第二儲存區域13◦儲存著第-儲存區域與第Γ 至第三快照存在於卷組中之位址。以下將以與上述 示上述所提到之各物件。 佩表 接著請參照「第2A 者透過LVM在卷組1〇〇 圖」本發明所提之方絲，當使用 中建立第-儲存_11〇後，執行有本 發明之快财切會在LVM上 快照儲存使用之邏輯卷，如「第3A .、4的儲存空間以建立供 儲存區域150，财象料_15“==卷為第三 之弟一快照121 (步驟21G)，另外，建㈣子=1〇 不僅侷限於第—儲存區域110建立± 的步驟並 由使用者設定欲建立第—儲麵域①如，也可以經 121 〇 '"一 之快照時建立第一快照 在第一快照121被建立之後，執杆古 第-快照t有本㈣之快照程式會將 _ 建第一快照121的第一時_ ^ 中的特定的第四儲存區域16Q，第四儲存區域160 知之儲存第一儲存區域彻之各快照索引訊息之儲存區 域’本發明所提之時職記錄有快·建謂間與快照的起始位 址，故第-時職161代表建立第一快照121的第一時間與第一 起始位址（步驟22G)。接著執行有本_之快照程式會設定第一 快照121的寫入前複製指標2〇〇為第一快照121之起始位址（步 騾_ 30)使得第快如J21進行寫入前複製時，被複製之資料區 塊會由寫入前複製指標2〇〇所指到之位址開始儲存（步驟24〇 )， 若資料區塊之大小為4KB，寫入前複製指標2〇〇指向第一快照 121之第一起始位址〇χ〇〇〇〇ΒΒ〇〇，當被複製之資料區塊（將被 更改之資籠塊）有3塊，則在寫人前複製完成之後，寫入前複 製指標200將會被修改為指向OXOOO〇bf〇〇的位址（步驟250)， 即移動了 12KB，上述之寫入前複製指標的修改過程僅為舉例， LVM配置的邏輯卷實際上是邏輯性的連續空間，也就是說，在物 1323414 理性的儲存裝置中並不一定為連續空間，但此部份為LVM之技術 手段’並非為本發明的重點，故不在此描述。 如上所述，使用者在建立第一快照121之後，對第一儲存區 域110所做之更改，包括刪除與修改，會被連續的儲存在第一快 ; 照121中，因此第一快照121會分為已儲存過資料區塊的第一已 使用區域121a以及尚未儲存資料區塊的第一未使用區域121b， 兩者以寫入前複製指標2〇〇所指向的位址為分界。 ❿ 當使用者在第二時間透過LVM建立第一儲存區域110的第二 快照時，執行有本發明的快照程式會以第二時間時寫入前複製指 標200所指向的位址為第二起始位址，在第三儲存區域彳5〇中， 接續在第一快照121之已使用區域121a之後，建立第二快照122 (步驟260) ’如「第3B圖」所示，其中第二快照122的儲存區 域會接續在第一快照121之已使用區域12ia之後。在第二快照 122建立之後，執行有本發明之快照程式會將第二快照up之第 •二起始位址，也就是寫入前複製指標2〇〇指向的位址，與第二時 •間儲存至第四儲存區域160中，成為第二時職162(步驟27〇> 若錢者再此之後更改第-贿輯㈣，雜更改的資料區境 會由寫入前複製指標200所指向的位址開始，依序儲存資料區塊 至第二快照122中。 同樣的，若使用者建立第一儲存區域11〇之第三快照123， 則第二快照123之第三已使用區域會接續在第二快昭122之第二 已使用區域i22a的後方，與第二快照122之第二未使用區: 122b重疊，並以寫入前複製指標2〇〇分隔第三已使用區域與第三 1323414 未使用區域，第四儲存區域160中也會儲存有第三快照的第 三時間戳163。如此，本發明即可以解決先前技術中所提之儲存 空間被浪費的問題。 若使用者欲取回舊資料，執行有本發明之快照程式會至第四 儲存區域160取得各快照的時間戳，若使用者欲取回第一快照的 舊資料，則執行有本發明之快照程式會由第一時間戳161中取出 第一快照121的第一起始位址，接著便可以索引至第一快照’μ •並取回舊資料，由於本發明並沒有改變寫入前複製之技術，因此 在快照中取回舊資料的方式與習知之方式相同，故不在此多做描 述。 若使用者欲刪除舊有的快照，由於依賴型快照只能由最先建 立的快照開始刪除，故在上述的實施例中，使用者必須先刪除第 一快照121 ’如「第2B圖」所示之流程，刪除快照時，執行有本 發明之快照程式會將第一時間戳161由第四儲存區域16〇中刪除 •(步驟31〇)，使得第四儲存區域僅剩下第二時間戳162與第三時 ' 間戳163 ’並至第二儲存區域130中，將第三儲存區域150的位 - 址修改為第二時間戳162中所記錄的第二起始位址，也就是第二 决知122的起始位址（步驟320)，如「第4圖」所示，如此便完 成刪_步驟，上述之兩步驟並沒有先後順序喊行關係，所以 刪除的步驟也可以先進行步驟32〇後再進行步驟31〇。 寫入前複製之技術並非只有在資料區塊被更改前，將被更改 之資料區塊複製到快照中一種，也有其他的技術，例如將更改後 的貧料區塊寫至快照中之技術，但不同的寫入前複製技術並不影 11 響本發明之實現。 再者’本發明亦可在 件散佈於軒互連腦射方式實減以不同元 私糸统的分散方式實現。 雖:、、、、本發明以前述之較电 定本發明，i "只施例揭露如上，然其並非用以限 疋本發明，任何抑 内，所為之更動^ 在不麟树明之精神和範圍 明之專梅護範_視以日日♦之專梅顿圍’因此本發 .準。 、η月田所附之申請專利範圍所界定者為 【圖式簡單說明】 ^圖係習知之快照分布於齡媒體中之示意圖。 、第2Α lih本發明所提之集中建立依賴型快照之方法之建』 方法流程圖。 第2B _本發明所提之集中建立依賴型快照之方法之· 方法流程圖。For the Volume Group (VG), the user can specify the storage area of the specific cache space in the volume group. LVM divides it into a logical volume (L〇gjca Volume; LV), and there is one in the volume group. The specific storage area is staggered with the address in each logical group. As shown in FIG. 1, the storage space is the volume group 'the first storage area 110, the first snapshot 121, and the second snapshot 122. The third snapshot 123 is a logical volume ' and the second storage area 13 stores an address in which the first storage area and the third to third snapshots exist in the volume group. The items mentioned above will be described below with the above. Please refer to "Part 2A through LVM in Volume Group 1". The square wire proposed by the present invention, when the first storage_11 is established in use, the fast cut with the present invention is executed in the LVM. The upper snapshot stores the logical volume used, such as "the storage space of the 3A., 4 to establish the storage area 150, the financial material _15" == the volume is the third brother of a snapshot 121 (step 21G), in addition, built (4) Sub=1〇 is not limited to the step of establishing the ±th portion of the first storage area 110 and is set by the user to establish the first storage area 1 as well as the first snapshot in the first snapshot. After a snapshot 121 is created, the snapshot program of the (4) snapshot will have a specific fourth storage area 16Q in the first time _ ^ of the first snapshot 121, and the fourth storage area 160 is known. The first storage area is stored in the storage area of each snapshot index message. The current time record of the present invention has a quick start address and a start address of the snapshot, so the first-time job 161 represents the first snapshot 121. The first time and the first start address (step 22G). Then, the snapshot program of the first snapshot 121 is set to set the pre-write copy index 2 of the first snapshot 121 to the start address of the first snapshot 121 (step _ 30) so that the fast copy is performed before the J21 is written. The copied data block will be stored by the address indicated by the pre-write copy indicator 2〇〇 (step 24〇). If the size of the data block is 4KB, the pre-write copy indicator 2 points to the first After the first start address of the snapshot 121, when the copied data block (the block to be changed) has 3 blocks, the copy is completed before the write is completed, before the write The copy indicator 200 will be modified to point to the address of OXOOO〇bf〇〇 (step 250), that is, 12 KB is moved. The modification process of the above-mentioned pre-write copy indicator is only an example, and the logical volume of the LVM configuration is actually logic. The continuous space of the character, that is to say, is not necessarily a continuous space in the rational storage device of the object 1323414, but the technical means of this part is not the focus of the invention, and therefore is not described here. As described above, after the first snapshot 121 is created, the changes made to the first storage area 110 by the user, including deletion and modification, are continuously stored in the first fast image 121, so the first snapshot 121 will be It is divided into a first used area 121a in which the data block has been stored and a first unused area 121b in which the data block has not been stored, and both are demarcated by the address pointed to by the pre-write copy index 2〇〇. ❿ When the user creates a second snapshot of the first storage area 110 through the LVM at a second time, executing the snapshot program of the present invention will use the address pointed to by the pre-write copy indicator 200 at the second time as the second The first address, in the third storage area ,5〇, after the used area 121a of the first snapshot 121, establishes a second snapshot 122 (step 260) 'as shown in FIG. 3B, where the second snapshot The storage area of 122 will continue after the used area 12ia of the first snapshot 121. After the second snapshot 122 is established, executing the snapshot program of the present invention will place the second start address of the second snapshot up, that is, the address pointed to by the pre-write copy indicator 2, and the second time. Stored in the fourth storage area 160 to become the second time position 162 (step 27〇> If the money person changes the first-bribe (4) again, the data area of the miscellaneous change will be copied by the pre-write copy indicator 200. The pointed address begins, and the data block is sequentially stored in the second snapshot 122. Similarly, if the user establishes the third snapshot 123 of the first storage area 11 , the third used area of the second snapshot 123 Continuing to be behind the second used area i22a of the second fast 121, overlapping with the second unused area: 122b of the second snapshot 122, and separating the third used area with the pre-write copy indicator 2〇〇 The third 1323414 unused area, the third time stamp 163 of the third snapshot is also stored in the fourth storage area 160. Thus, the present invention can solve the problem that the storage space mentioned in the prior art is wasted. Retrieve the old data and execute the invention The program will obtain the timestamp of each snapshot in the fourth storage area 160. If the user wants to retrieve the old data of the first snapshot, the snapshot program executing the present invention will take the first snapshot 121 from the first timestamp 161. The first start address can then be indexed to the first snapshot 'μ• and retrieve the old data. Since the present invention does not change the technique of pre-write copying, the way to retrieve the old data in the snapshot is in a conventional manner. The same is true, so it is not described here. If the user wants to delete the old snapshot, since the dependent snapshot can only be deleted by the first snapshot, in the above embodiment, the user must delete the first snapshot first. 121 ' As shown in the "Picture 2B", when the snapshot is deleted, executing the snapshot program of the present invention deletes the first time stamp 161 from the fourth storage area 16 (step 31), so that the fourth Only the second timestamp 162 and the third time 'stamp 163' remain in the storage area and into the second storage area 130, and the bit-address of the third storage area 150 is modified to be recorded in the second timestamp 162. Second start address That is, the start address of the second decision 122 (step 320), as shown in "Fig. 4", so that the deletion step is completed, and the above two steps do not have a sequential relationship, so the step of deleting You can also perform step 32 before proceeding to step 31. The technique of copying before writing is not only to copy the changed data block to the snapshot before the data block is changed, but also other techniques, such as The changed poor block is written to the technology in the snapshot, but different pre-write copying techniques do not affect the implementation of the present invention. Furthermore, the present invention can also be spread in the inter-connected brain It is achieved by the method of decentralization of different private systems. Although the invention is based on the above-mentioned invention, i "only the examples are disclosed above, but it is not intended to limit the invention, In addition to the spirit and scope of the Lin Shuming, the beauty of the Mei _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ The definition of the patent application scope attached to η月田 is [Simple description of the schema] ^The schematic diagram of the snapshot of the schema is distributed in the age media. 2nd lih The flow chart of the method for constructing a method for establishing a dependent snapshot in the present invention. 2B - A flow chart of a method for constructing a dependent snapshot in the present invention.
第3A圖係本發明實施例所提之建立快照過程示意圖。 第3B圖係本發明實施例所提之建立快照過程示意圖。 第4圖係本發明實施例所提之刪除快照過程示意圖。 【主要元件符號說明】 100 儲存空間 110 第一儲存區域 121 第一快照 121a 第一已使用區域 121b 第一未使用區域 12 122 第二快照 122a 第二已使用區域 122b 第二未使用區域 123 第三快照 123a 第三已使用區域 123b 第三未使用區域 130 第二儲存區域 140 寫入前複製指標 150 第三儲存區域 160 第四儲存區域 161 第一時間戳 162 第二時間戳 163 第三時間戳 1323414 步驟210建立第一快照 步驟220儲存第一快照之第一時間戳 步驟230設定寫入前複製指標為第-快照之起始位址 步驟240寫入前複製 步驟250 |改寫入前複製指標為可繼續儲存區塊之位址 步驟260卩寫人前複製指標騎始位址建立第二快照 步驟270儲存第二快照之第二時間戳 步驟310刪除第一時間戳 v驟320叹定儲存快照之儲存區域之起始位址為第二快照 之起始位址 13FIG. 3A is a schematic diagram of a process of establishing a snapshot according to an embodiment of the present invention. FIG. 3B is a schematic diagram of a snapshot process established by the embodiment of the present invention. FIG. 4 is a schematic diagram of a process of deleting a snapshot mentioned in the embodiment of the present invention. [Main component symbol description] 100 storage space 110 first storage area 121 first snapshot 121a first used area 121b first unused area 12 122 second snapshot 122a second used area 122b second unused area 123 third Snapshot 123a third used area 123b third unused area 130 second storage area 140 pre-write copy indicator 150 third storage area 160 fourth storage area 161 first time stamp 162 second time stamp 163 third time stamp 1323414 Step 210: Establish a first snapshot step 220 to store a first timestamp of the first snapshot. Step 230: Set a pre-write copy indicator to a start address of the first snapshot. Step 240: Write before copying step 250 | Continuing to store the address of the block 260 卩 人 复制 复制 复制 骑 骑 骑 骑 骑 建立 建立 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 270 The starting address is the starting address of the second snapshot 13