201116993 六、發明說明： 【發明所屬之技術領域】 特別是關於一 本發明係關於一種伺服器的壓力測試的方法 種可動態調整施加之磨力的塵力測試方法。 【先前技術】 隨著電普及與麟技觸快速發展，經_路可以快速 的獲取所㈣資訊，更可以制多樣㈣服務。也由於應用於網 路的技術越來越進步，僅由普通電腦或設備所能提供的服務以不 敷使用’故發展出了舰n的技術，職係為—種善於處理網 路技術之電腦平台，其可連結至各種網路系統，諸如網際網路 (Internet) ^ ^^^.P#^(intranet). M^fa1>,^(extranet) . 路(Local Area Network ’ LAN)。伺服器並對透過網路系統相連結 的電腦提供各種的應用服務。 為了確保词服器於提供服務時的穩定性，壓力測試被列為飼 服器出被出貨前非常f要的職項目之…舰器的壓力測試係 可針對舰H的各個f要元件如中央處理器或網路卡進行測試， 以確認伺服器是否可在—定的壓力下正常地運作。 於傳統的壓力測試方法，測試人員依據經驗配置測試用腳 本’並可令職程序依據腳本對特定的硬體連續進行長時間的運 轉測試。-次的壓力測試可能長達72小時，且一旦腳本配置完成 並據以進制試，不論運行過程巾制試的硬體實際上接受到的 壓力或負載情況如何’測试程序都會按照腳本的設定進行直到測 201116993 試結束。但㉔職力以及硬體負載可能隨著時間或舰器實際 運作狀況變化，崎生_職施加_力翻_符的情況。 來測試人員即使在測試過程中發現需要修改虔力配 置’也只料待此次測試結束或是直接中_m，再設計下-次 測試所需要_本。的碰腳本難可能甚至需要數次，而 有浪費許多測試時間和修改腳本所需之功夫的問題。 此外，傳統的壓力測試方法若對飼服器施加過高的壓力，常 •會出現由於硬體負荷過重而引起系統崩潰，甚至是伺服器的硬體 ，受到損害的問題。且系統崩潰或硬體損毁時的測試數據無法即 時被保存，更增加了後續對問題追蹤時的麻煩。 【發明内容】 為了解決上述問題，本發明提供一種伺服器的壓力測試方 法。錬ϋ的壓力職方法係顧於對具有複數個硬體模組的一 値器施加-塵力以進行壓力測試，而伺服器電性相連於一饲服 器管理系統。舰㈣壓力測試方法包括執行—職程序，其中 測試程序包括··接收-測試腳本，其中測試腳本指定一指定測試 壓力予至少一個硬體模組；依據指定測試壓力對被指定的硬體模 組施加壓力；藉由伺鮮管理紐讀取至少—個硬雜組的一模 組狀態；分析模組狀態，並得到一現行測試昼力；判斷現行測試 壓力是否與指定職壓力補；以及當不相符時，依據現行測試 壓力與指定測賴力調整對被指定的硬體模組施加的麗力。 祠服器的硬體模組可為一中央處理器、一記憶體、一硬碟或 201116993 是-電源供應器。而硬體模組的模組狀態可為硬體模組的一硬體 溫度、-硬觀雜速或是一硬體神。另外，模組狀態亦可為 硬體模組的一狀態暫存器。 _ 根據本發明之-實施範例，硬體模組可包括至少二子模組。 則測試程序的「讀取鑛㈣至少―個硬體模組的模組狀態」的 步驟’係讀取硬體模組的這些子模組的硬體狀態。且測試程序的 「依據現行測試壓力與贼測試動調整對被指定的硬體模組施 加的壓力」时驟，係調整對被指定的硬體模組的這些子模組施 加的壓力。 根據本發明之另-實施制，舰_壓力測試方法另可包 括執行-預警程序。預警程序可包括：對至少一個硬體模組界定 -壓力警戒值’讀取對舰力警纽之硬顏組賴組狀態丨分 析對應勤警戒值之模域態，並制賴壓力警纽之現行測 試壓力；躺對應壓力縣值之現行測試壓力是否大於壓力警戒 值；以及當對賴力警紐之現行職壓力大於壓力警戒值時， 發送一警戒信號予測試程序。 而測試程序另可包括：接收警戒信號；以及調降對現行測試 壓力大於壓力警戒值的硬體模組施加的壓力。 根據本發明之.又一實施範例，預警程序另可包括：對至少一 個硬體模組界定一壓力臨界值；讀取對應壓力臨界值之硬體模組 的模組狀態；分析對應壓力臨界值之模組狀態，並得到對應壓力 臨界值之現行測試壓力；判斷對應壓力臨界值之現行測試壓力是 201116993 否大於壓力臨界值；以及當對應壓力臨界值之現行測試壓力大於 壓力臨界值時，發送一臨界信號予測試程序。 而測試程序則另可包括：接收臨界信號；以及中斷測試程序。 此外，上述之伺服器的壓力測試方法的預警程序可與測試程 序同時執行。伺服器的壓力測試方法則可運行於伺服器或伺服器 遠端管理系統，且値n管理系祕運行於遠端(rem〇te)。 絲上所述，根據本發明之實施範例之伺服器的壓力測試方法 籲可依據硬體模組的現行測試壓力動態調整施加於指定的硬體模組 的[力’以使其符合所冑的指定測試壓力。且伺服器的壓力測試 方法藉由鮮鱗監控硬體顧。當硬贿_行職壓力高於 壓力警戒值或是壓力臨界值時，預警程序可令職程序適當地調 降壓力’或是備份測試資料。 因此伺服器的壓力測試方法可不需中斷測試程序，而免除許 多不符合職且無謂的戦咖和錢腳本的功夫。健器的壓 ’力測試方法並可有效地在進行壓力測試的過程中保護飼服器，若 於測試過轉生意外情形’亦可提供備份制試資料以供測試人 員。 【實施方式】 々以下在實施方式中詳細敘述本發明之詳細特徵以及優點，其 =谷足以使任何熟習_技藝者了解本發明之技術内容並據以實 把且根據本說明書所揭露之内容、申請專利範圍及圖式，任何 熟習相關技藝者可㈣地理解本發明_之目的及優點。 201116993 本發明提供一種伺服器的壓力測試方法，其適用於對具有複 數個硬體模組的一飼服器施加一麼力以進行壓力測試，且飼服器 電性相連於一伺服器管理系統。 请參考「第1A圖」，其係為根據本發明一實施範例之伺服器 方塊示思圖。如「第1A圖」所繪示，伺服器2〇包含複數個硬體 模組’分別為一中央處理器(Central Processing Unit，CPU)22、一 北橋晶片23、一南橋晶片24、一記憶體26、一硬碟27以及一電 源供應盗28。伺服器20並透過内部的一基板管理控制器 (Baseboard Management Controller，BMC)25 與伺服器管理系統 30 電性相連。 伺服器20可使用Unix的Linux、FreeBSD或是微軟(Micr〇s〇ft) 的Windows Server 2003等系統，亦可為磁碟作業系統(1^企 Opemtmg System ’ DOS)或是可延伸韌體介面(Extens]ible Firmware201116993 VI. Description of the Invention: [Technical Field to Which the Invention Is Applicable] In particular, the present invention relates to a method for testing a pressure of a servo, which is a method for dynamically adjusting the applied dust force. [Prior Art] With the rapid development of electricity and the rapid development of the technology, the _ Road can quickly acquire the information (4), and can also make a variety of services. Also, because the technology applied to the network is getting more and more advanced, the service that can be provided by ordinary computers or equipment is not enough to use the technology of the ship, and the grade is a computer that is good at processing network technology. A platform that can be connected to various network systems, such as the Internet (^)^^^.P#^(intranet). M^fa1>, ^(extranet). Road (Local Area Network 'LAN). The server provides various application services to computers connected to the network system. In order to ensure the stability of the word processor in providing service, the stress test is listed as a very important job before the shipment is shipped. The pressure test of the ship can be used for each component of the ship H. The central processor or network card is tested to verify that the server is functioning properly at a given pressure. In the traditional stress test method, the tester configures the test script based on experience, and can use the script to perform continuous long-term operation tests on specific hardware according to the script. - The stress test may be as long as 72 hours, and once the script is configured and tested according to the hex, regardless of the pressure or load condition actually received by the hardware of the running process, the test program will follow the script. The setting is carried out until the end of the test 201116993. However, the 24th position and the hardware load may change with time or the actual operating conditions of the ship. The tester even needs to modify the force configuration during the test process, and only needs to wait for the end of the test or directly in the _m, and then design the next test. It's difficult to touch a script even several times, and there is a problem that wastes a lot of testing time and the effort required to modify the script. In addition, if the traditional stress test method exerts excessive pressure on the feeding device, there is often a problem that the system is collapsed due to excessive hardware load, and even the hardware of the server is damaged. The test data when the system crashes or the hardware is damaged cannot be saved immediately, which increases the trouble of following the problem tracking. SUMMARY OF THE INVENTION In order to solve the above problems, the present invention provides a pressure testing method for a server. The pressure method of the 錬ϋ is to apply a dust force to a tamper with a plurality of hardware modules for stress testing, and the servo is electrically connected to a feeding management system. The ship (four) stress test method includes an execution-service program, wherein the test program includes a receive-test script, wherein the test script specifies a specified test pressure to at least one hardware module; the designated hardware module is specified according to the specified test pressure Apply pressure; read at least one module state of the hard group by the servo management button; analyze the module status and obtain an active test force; determine whether the current test pressure is supplemented with the specified job pressure; When matching, adjust the power applied to the specified hardware module according to the current test pressure and the specified measurement force. The hardware module of the server can be a central processing unit, a memory, a hard disk or a 201116993-power supply. The module state of the hardware module can be a hardware temperature of the hardware module, a hard view speed or a hard body. In addition, the module state can also be a state register of the hardware module. According to an embodiment of the present invention, the hardware module may include at least two sub-modules. Then, the "step of reading the mine (four) at least one module state of the hardware module" of the test program reads the hardware state of the submodules of the hardware module. And the test program "adjusts the pressure applied to the specified hardware module according to the current test pressure and the thief test adjustment", and adjusts the pressure applied to the sub-modules of the designated hardware module. According to another embodiment of the present invention, the ship_pressure test method may further include an execution-warning procedure. The early warning procedure may include: defining at least one hardware module-pressure warning value' reading the state of the hard-group group of the ship's force alarm, analyzing the mode domain state corresponding to the service alert value, and relying on the pressure police Current test pressure; whether the current test pressure corresponding to the pressure county value is greater than the pressure alert value; and when the current job pressure of the Lai Li police is greater than the pressure alert value, a warning signal is sent to the test procedure. The test procedure may further include: receiving a warning signal; and reducing the pressure applied to the hardware module whose current test pressure is greater than the pressure warning value. According to still another embodiment of the present invention, the early warning program may further include: defining a pressure threshold for the at least one hardware module; reading a module state of the hardware module corresponding to the pressure threshold; and analyzing the corresponding pressure threshold The module state and the current test pressure corresponding to the pressure threshold; the current test pressure for determining the corresponding pressure threshold is 201116993 is greater than the pressure threshold; and when the current test pressure corresponding to the pressure threshold is greater than the pressure threshold, A critical signal is given to the test program. The test procedure may further include: receiving a critical signal; and interrupting the test procedure. In addition, the warning program of the above-mentioned server stress test method can be executed simultaneously with the test program. The server's stress test method can be run on the server or server remote management system, and the 管理n management system runs at the remote end (rem〇te). As described above, the pressure test method of the server according to the embodiment of the present invention can dynamically adjust the [force] applied to the specified hardware module according to the current test pressure of the hardware module to make it conform to the Specify test pressure. And the server's stress test method is monitored by fresh scales. When the hard bribery _ job pressure is higher than the stress alert value or the pressure threshold, the early warning procedure can make the procedure properly reduce the pressure ‘ or back up the test data. Therefore, the server's stress test method eliminates the need to interrupt the test procedure, and eliminates the incompetent and unnecessary use of the coffee and money scripts. The pressure test method of the health device can effectively protect the feeding device during the stress test, and can also provide backup test data for the tester if the test has been carried out. [Embodiment] The detailed features and advantages of the present invention are described in detail below in the embodiments, which are sufficient to enable any skilled artisan to understand the technical contents of the present invention and to present the contents according to the present disclosure. To the extent of the patent application and the drawings, any person skilled in the art can understand the purpose and advantages of the present invention. 201116993 The present invention provides a pressure testing method for a server, which is suitable for applying a force to a feeding device having a plurality of hardware modules for stress testing, and the feeding device is electrically connected to a server management system. . Please refer to "FIG. 1A", which is a block diagram of a server block according to an embodiment of the present invention. As shown in FIG. 1A, the server 2 includes a plurality of hardware modules 'a central processing unit (CPU) 22, a north bridge chip 23, a south bridge chip 24, and a memory. 26. A hard disk 27 and a power supply pirate 28. The server 20 is electrically connected to the server management system 30 through an internal Baseboard Management Controller (BMC) 25. The server 20 can use Unix Linux, FreeBSD or Microsoft (Micr〇s〇ft) Windows Server 2003 and other systems, and can also be a disk operating system (1^Enterprise Opemtmg System 'DOS) or an extendable firmware interface. (Extens]ible Firmware
Interface，EFI)的系統。且伺服器2〇亦可各種廠牌之各種伺服器 產品，本發明並不對其限制。 中央處理器22透過北橋晶片23與伺服器2〇之主機板(未繪示） 上的高速裝置電性相連。例如中央處理器22可透過北橋晶片23 存取記憶體26,亦需透過北橋晶片23存取南橋晶片24以進一步 控制與南橋晶片24相連的其他硬體模組。南橋晶片24與主機板 上的其他硬體模組相連，以將伺服器2〇中運轉速度較慢的硬體模 組提供給中央處理器22。舉例而言，中央處理器22可透過與南橋 晶片24相連之序列先進技術附件（Serial Advanced TedmQk>gyInterface, EFI) system. Moreover, the server 2 can also be used for various server products of various brands, and the present invention is not limited thereto. The central processing unit 22 is electrically connected to the high speed device on the motherboard (not shown) of the server 2 through the north bridge chip 23. For example, the central processing unit 22 can access the memory 26 through the north bridge chip 23, and also access the south bridge wafer 24 through the north bridge wafer 23 to further control other hardware modules connected to the south bridge wafer 24. The south bridge chip 24 is coupled to other hardware modules on the motherboard to provide a hard-working hardware module in the server 2 to the central processor 22. For example, the central processing unit 22 can access the serial advanced technology attachment (Serial Advanced TedmQk> gy) connected to the south bridge chip 24.
Attachment ’ Serial ΑΤΑ ’亦簡稱SATA)控制器存取硬碟27，亦 201116993 可透過與南橋晶# 24相連之通用序顏流排(Univer>saI Serial Bus ’ USB)控制器或周邊組件互連（pe^^ c〇mp_tAttachment ' Serial ΑΤΑ ' is also referred to as SATA) controller access to hard disk 27, and 201116993 can be interconnected via a universal sequencer (Univer > saI Serial Bus 'USB) controller or peripheral components connected to South Bridge #24 ( Pe^^ c〇mp_t
InterC〇nneCt，PCI)控制器再與其他的硬體模組相連。伺服器20的 電源供應器28可連接到主機板以及硬碟27，而中央處理器22等 配置於主機板上的硬體模組再經由主機板獲得電力。 伺服器管理系統30亦被稱之為智能平台管理介面(IntemgentThe InterC〇nneCt, PCI) controller is connected to other hardware modules. The power supply 28 of the server 20 can be connected to the motherboard and the hard disk 27, and the hardware module disposed on the motherboard, such as the central processing unit 22, can obtain power via the motherboard. The server management system 30 is also referred to as an intelligent platform management interface (Intemgent).
Platform Management Interface ’ IMPI) ’ 其係能透過 BMC 25 對伺 服器20提供遠端監視、管理、錯誤診斷和系統恢復等功能。伺服 器官理系統3G可藉由各種的界面來管_服器2()，並可透過網路 或序列埠來通知管理人員。 一般來說，伺服器管理系統30提供智能平台管理匯流排 (Intelligent Platform Management Bus，IPMB)和 BMC 25 溝通。舉 例而言，伺服器管理系統30可以是惠普(Hp)公司的iL〇系統、戴 爾DELL公司的iDRAC***，或是英特爾(Intel)公司的ESB2系 統。 BMC 25又可稱為管理晶片，能夠透過配置於伺服器之各 硬體模組上的偵測器(sensor)監控各個硬體模組的運行狀況。BMC 25就像是一***立的電腦，具有自己内建的處理器與記憶模組等 資源。BMC 25在伺服器20未開機的情況下便可運作，且bmc 25 的運作居使用自己具有的資源·，而不會佔用伺服器2〇之硬體模組 的其他資源。 藉由上述硬體結構，根據本發明之一實施範例之伺服器的壓 力測試方法執行一測試程序。請參照r第2圖」，其係為根據本發 201116993 明一實施範例之測試程序之流程示意圖。由「第2圖」可以見悉5 測试程序可包括步驟S40 :接收一測試腳本，其中測試腳本指定一 指定測試壓力予祠服器的至少一個硬體模組；步驟⑷：依據指定 測試壓力對被指定的硬聽模組施加壓力；步驟S42 :藉由伺服器管 理系統讀取伺服器的至少一個硬體模組的一模組狀態；步驟S43 : 分析模組狀態，並得到一現行測試壓力；步驟S44 :判斷現行測試 壓力是否與指定測試壓力相符；步驟S4L依據現行測試壓力與指 定測試壓力調整對被指定的硬體模組施加的壓力；以及步驟S46 : 對被指定的硬體模組施加原本的壓力。 於步驟S40測試程序接收測試腳本。一測試人員可藉由測試 腳本對至彡、-個硬難組(如+域理_ 22或記憶體26)指定所需 的指定測試壓力，其巾指定戦壓力的指定方式可例如以百分比 的方式指定。舉例來說’峨腳本的内容可以是「cpu 6G%，記 憶體80%」。 測試权序接收測試腳本後，於步驟S41依據指定壓力對被指 定的硬體模組施加壓力，試程序可依據指定壓力令中央處理器 22不斷執行-定數量的執行绪(thread)，亦可對記憶體％與硬碟 27進行一定頻率或次數的存取以作為測試的壓力。 •測試程序於步驟S42由BMC 25可以讀取各個硬體模組的模 組狀態。其中模組狀態可為由硬體模組上的制器獲得的一硬體 溫度、-硬體風祕速或是-硬體功率；亦可為硬體模組的一狀 ，暫存flatus regis㈣。更詳_說，模組狀態可以是中央處理 态22或是記憶體26的狀態暫存器、風扇轉速或是溫度，亦可以 201116993 是電源供應ϋ 28的狀態暫存n或是功率。 &接著於步驟S43 ’測試程序分析由BMC 25所讀取的模組狀 ，並得到現行測試壓力。由於在測試的過程中，真正施加予硬 體模、、且的[力以及硬體模级的負載可能隨著時間或飼服器如運作 而變化，因此被指定的硬體模組所受到的墨力並不一定與測試腳 本中的指定測試壓力相同。 2例而言’下述職均可能造成現行職壓力與指定測試壓 #力不符：伺服器20的各個硬體模組實_置情況；測試程序或其 他伺服器20内的程序被運行時，硬體模組的使用情況；壓力測試 被執行的時間長短；以及執行壓力測試時，當前之硬體模組的運 行情況。 更詳細地武’飼服器2〇的各個硬體模組實際配置情況，可包 括中央處㈣22的數;t;記憶體％的容量；歧硬碟27的容量。 测試程序或其他伺服器20内的程序被運行時，硬體模組的使用情 鲁况則可包括記憶體26或硬碟27剩餘的可用容量。而執行壓力測 試時，當前之硬體模組的運行情況可包括：#前同時運行且佔用 +央處理器22之測試項目的數量；或是當前同時運行且佔用一輸 入輪出單元(Input/Outputunit，未標示)之測試項目的數量。 於步驟S44，測試程序便判斷現行測試壓力是否與指定測試壓 力柄符。步驟S44可躺現行測試測試壓力是否與指定測試壓力 相同’或是在-可容許的誤差範圍内(例如加減3%) ^ ^現行測試 堡讀指定測試壓力不相符，於步驟泌中，測試程序依據現行 測4壓力與指定測試麗力，調整對被指定的硬體模組施加的壓 11 201116993 力也仇疋成，測試程序能依據現行測試壓丈動態地加強或減輕 e加於被;日々的硬體模組的麼力，以使其符合測試腳本中的指定 測試壓力。 、而若現行測試壓力與指定測試壓力相符，則於步驟S46中測 «式程序繼_對被指定的硬體模組施加目前的壓力。此外，測試程 序並可料執行步驟sc至步驟MG，以更靈活地動態調整所需 施加的屋力。 接下來請同時參照「第2圖」與「第1B圖」，其中「第m 圖」係為根據本發明另—實施範例之鑛器方塊示賴。如「第· 1B圖」所緣示，_央處理器22具有一中央處理單元22i以及222。 於此實施範例中，被指定的硬體模組可以包括至少兩個子模组， 例如中央處理器22之中央處理單元221以及從。類似地，記憶 體26亦可具衫個子記憶體，或是硬碟27可具有多個子硬碟(例 如硬碟27的多個槽）。 測試人員 H則試腳本中對硬體模組的各個子模，組給予盆指 定測試屋力，例如分別對中央處理單元221與222要求卿。與· .的指定測試勤。則於步驟S42 t，測試程序可藉由赚^讀 取各個子拉組的模組狀態.，例如中央處理單元221以及迎的田 度。且測試程序並可依擄各個子模組的模組狀態分析得到其個$ 的現打測倾力，並馳指定的硬體模組之多個子模組調整需施 加的壓力。 的壓力測試方法之測試程序 ’亦可進行模組内部(之各個 換句話說，本發明提供的伺服器 不但可進行各硬體模組間的壓力調節 12 201116993 子模組間）的壓力調節。 根據本發明之—實絲例，為了麵再進行壓力職的過程 中，飼服器20的硬體模組出現不堪負荷而崩潰甚至是燒毀的情 形，飼服器的壓力測試方法另可具有一預警程序。 明多…、第3圖」’其係為根據本發明一實施範例之預警程序 之飢程示w圖由第3圖」可以見悉，預警程序可包括步驟85〇: 對至少-個硬雜組界定—壓力警戒值；步驟如：讀取對應壓力 籲警戒值之硬體模組的模組狀態；步驟说：分析對應壓力警戒值之 杈組狀‘癌，並得到對應壓力警戒值之現行測試壓力；步驟的：判 斷對應壓力警戒值之現行測試壓力是否大於壓力警戒值；以及步 驟S54 :發送一警戒信號予測試程序。 於步驟S50’預警程序可指定壓力警戒值給至少一個硬體模 、、且舉例而5，若電源供應器28所發出的功率過高，可能會燒毀 整個飼服器20,因此應對其進行監控。接著預警程序可於步驟⑹ 藉由BMC 25讀取對應壓力警戒值之硬體模組的模組狀態；並於 步驟S52分析此模組狀態，以得到對應的現行測試壓力。更詳細 的5兒’本5兒明書述之「對應壓力警戒值的模組狀態/現行測試壓力」 係指被界定有壓力警戒值的硬體模組的模組狀態/現行測試壓力_。 也就疋说，預警程序可定時確認界定有壓力警戒值之硬體模組的 模組狀態，並分析得到界定有壓力警戒值之硬體模組的現行測試 壓力。 接著於步驟S53’預警程序判斷對應壓力警戒值之硬體模組之 .現行測試壓力是否大於壓力警戒值。若此硬體模組的現行測試壓 13 201116993 力已大於壓力警戒偟， 試程式。 預警程序便於步驟發送警戒信號予測 相對地，測試料另可包細下步驟：接收 調降對現行測力大職力警献的硬體模纟辑杨祕力及 為了避免需[控的硬體模組(即被界定錢力警戒值的硬體模句 發生故障等_，戦鱗__錄之後，姆對此硬體 模組施加的勤。預鶴組亦可將目前戦鱗的執行狀況等相 關資訊進彳了備份，赌後續賴人貞的追縱參考。 此外，根據本發明另-實施範例，預警程序另可對需監控 硬體模組設定-壓力臨概，以更妥當地保護值器2g。請參照 第4圖」’其係為根據本發明另-實施範例之預警程序之流程示 %、圖由第4圖」可以見悉’預警程序另可包括步驟s6〇:對至 少-個硬體模組界定壓力臨界值；倾S61 :讀取對應壓力臨界值 之硬體模組的模組狀態；步驟S62 :分析對應壓力臨界值之模組狀 態，並得到對應壓力臨界值之現行測試壓力；步驟S63 :判斷對應 壓力臨界值之現行測試壓力是否大於壓力臨界值；以及步驟s64 : 發送一臨界信號予測試程序。 本實施範例與上述實施範例不同之處在於：預警程序另對硬 體模組指定壓力臨界值，且壓力臨界值比警戒值較高。舉例而言， 中央處理器的壓力警戒值可設為80%，壓力臨界值則可設為 9〇%。而測試程序則另可包括以下步驟：接收臨界信號；以及中斷 測試程序。 若現行測試壓力已高於壓力臨界值，表示伺服器20有很高的 201116993 機率發生系統崩潰或是硬體損毀的危險狀況。因此一旦測試程序 接收到臨界_，測試程賴暫蔣止顯。鮮料發送=界 信號後亦可進行資料的備份以提供測試人員參考。且若預馨程序 發現異常’柯於備份資料之後將伺服器2㈣閉，以確保=器 20的安全。 根據本剌之舰H的測試方法係可啊執行觀程序以及 預警程序。測試程序以及預警程序可同時藉由BMc Μ監控飼服 φ器20的硬體模組的模組狀態，並依據實際情況動態調整對硬體模 組(及其子模組)施加的壓力。伺服器的測試方 資源執行。 伺服器的測試方法亦可運行於伺服器管理系統3〇，且飼服器 管理系統30可運行於遠端。換句話說，遠端的飼服器管理系統扣 可透過區網或網際網路連接於伺服器2〇2BMC 25。而運行於伺 φ服态官理系統30的伺服器的壓力測試方法亦可透過BMC 25以及 網路遠端監控舰器20的硬體模組，並動態調整對於指定的硬體 模組所施加的壓力。 綜上所述，根據本發明之實施範例之伺服器的壓力測試方法 可:析恤器之硬體模組賴組狀態，並得到硬體模組的現行測 忒壓力，再據以動態調整施加於指定的硬體模組的壓力，以使其 符合所需的指定測試壓力。如此一來，伺服器的壓力測試方法可 不需中斷測試程序而夠隨著伺服器實際運行的狀況調整壓力，免 除許多因嘗試錯誤(try an(i error)而花費的測試時間和修改腳本的 15 201116993 功夫。 更進一步地’伺《賴力峨方轉由卿程序監控硬體 核組。當硬驗_行測賴力高補力警戒值或是勤臨界值 時，預警料可令峨料適t__力，或是備份峨倾。 因此預警程序可有效地在進行勤測試的過針保護词服器若 於測試過程發生意外情形，亦可提供備份的測試資料以供測試人 員分析、了解。 —雖然本發明以前述之較佳實施例揭露如上，然其並非用以限 定本發明’任何熟習相像技藝者，在賴離本發明之精神和範圍 内’當可作些許之更動烟飾，因此本發明之翻賴範圍須視 本說明書所附之申請專利範圍所界定者為準。 【圖式簡單說明】 第1A圖係為根據本發明一實施範例之伺服器方塊示意圖； 第1B圖係為根據本發明另一實施範例之伺服器方塊示意 囬 · _， 第2圖係為根據本發明一實施範例之測試程序之流程示意 rat! · 圍， 第3圖係為根據本發明一實施範例之預警程序之流程示意 圖；以及 第4圖係為根據本發明另一實施範例之預警程序之流程示意 圖。 【主要元件符號說明】 201116993 20 伺服器 22 中央處理器 221,222 中央處理單元 23 北橋晶片 24 南橋晶片 25 基板管理控制器 26 記憶體 φ 27 硬碟 28 電源供應器 30 伺服器管理系統The Platform Management Interface ’ IMPI) ’ enables the remote monitoring, management, error diagnosis, and system recovery of the server 20 through the BMC 25. The servo organ system 3G can control the server 2() through various interfaces, and can notify the administrator through the network or the serial port. In general, the server management system 30 provides intelligent platform management bus (IPMB) and BMC 25 communication. For example, the server management system 30 can be an HP (Hp) company's iL system, Dell DELL's iDRAC system, or Intel's ESB2 system. The BMC 25 can also be called a management chip, and can monitor the operation status of each hardware module through a sensor disposed on each hardware module of the server. The BMC 25 is like a standalone computer with its own built-in processor and memory modules. The BMC 25 operates when the server 20 is not powered on, and the bmc 25 operates using its own resources without consuming other resources of the server's hardware modules. With the above hardware structure, a test procedure for a servo pressure test according to an embodiment of the present invention is performed. Please refer to the second figure of r, which is a schematic flow chart of the test procedure according to the embodiment of the present invention. It can be seen from "Fig. 2" that the 5 test procedure can include the step S40: receiving a test script, wherein the test script specifies a specified test pressure to the at least one hardware module of the server; and step (4): according to the specified test pressure Applying pressure to the designated hard listening module; step S42: reading a module state of at least one hardware module of the server by the server management system; step S43: analyzing the module state, and obtaining an active test Pressure; step S44: determining whether the current test pressure is consistent with the specified test pressure; step S4L adjusting the pressure applied to the designated hardware module according to the current test pressure and the specified test pressure; and step S46: for the designated hard mold The group applies the original pressure. The test program receives the test script in step S40. A tester can specify the required test pressure to the hard disk group (such as + domain _ 22 or memory 26) by the test script, and the specified pressure of the towel can be specified, for example, in percentage. Method specified. For example, the content of the script can be "cpu 6G%, memory 80%". After the test order receives the test script, the pressure is applied to the designated hardware module according to the specified pressure in step S41, and the test program can continuously execute the predetermined number of threads according to the specified pressure. A certain frequency or number of accesses to the memory % and the hard disk 27 are used as the pressure for the test. • The test program can read the module state of each hardware module by the BMC 25 in step S42. The module state can be a hardware temperature obtained by the controller on the hardware module, - hard wind speed or - hard power; can also be a shape of the hardware module, temporary flatus regis (four) . More specifically, the module status can be the central processing state 22 or the state register of the memory 26, the fan speed or the temperature, or the 201116993 is the state of the power supply ϋ 28 to temporarily store n or power. & Next, in step S43', the test program analyzes the module shape read by the BMC 25 and obtains the current test pressure. Since the load applied to the hard mold during the test, and the load of the hard mold stage may change with time or the feeder as the operation, the specified hardware module is subjected to The ink force is not necessarily the same as the specified test pressure in the test script. In the case of 2 cases, the following jobs may cause the current job pressure to be different from the specified test pressure: the hardware modules of the server 20 are in the actual state; when the test program or other programs in the server 20 are run, the hard The use of the body module; the length of time the stress test was performed; and the operation of the current hardware module when performing the stress test. More specifically, the actual configuration of each hardware module of the ’ 饲 器 , , , can include the number of the central (four) 22; t; the capacity of the memory%; the capacity of the hard disk 27. When the test program or other programs in the server 20 are run, the usage of the hardware module may include the remaining available capacity of the memory 26 or the hard disk 27. When the stress test is performed, the current operation of the hardware module may include: ## before running and occupying the number of test items of the central processor 22; or currently running at the same time and occupying an input round-out unit (Input/ The number of test items for Outputunit, not shown). In step S44, the test program determines whether the current test pressure is associated with the specified test pressure. Step S44 can be used to test whether the test pressure is the same as the specified test pressure 'or within the allowable error range (for example, plus or minus 3%) ^ ^ The current test test reads the specified test pressure does not match, in the step of the test, the test procedure According to the current test 4 pressure and the specified test Lili, adjust the pressure applied to the designated hardware module. 11 201116993 The force is also arrogant, the test program can dynamically strengthen or reduce the e-addition according to the current test pressure; The power of the hardware module to match the specified test pressure in the test script. If the current test pressure is consistent with the specified test pressure, then the step (step S46) is followed by applying the current pressure to the designated hardware module. In addition, the test program can perform step sc to step MG to more flexibly adjust the required house force dynamically. Next, please refer to "Fig. 2" and "Fig. 1B" at the same time, wherein "mth diagram" is a mine block according to another embodiment of the present invention. The "Central Processing Unit 22" has a central processing unit 22i and 222 as shown in the "FIG. 1B". In this embodiment, the designated hardware module may include at least two sub-modules, such as the central processing unit 221 of the central processing unit 22 and the slave. Similarly, the memory 26 may also have a shirt memory, or the hard disk 27 may have a plurality of sub-hard disks (e.g., a plurality of slots of the hard disk 27). The tester H tests the script for each submodule of the hardware module, and the group gives the basin the test house force, for example, the central processing units 221 and 222 respectively. Designated test with . Then, in step S42 t, the test program can read the module status of each sub-group by earning ^, for example, the central processing unit 221 and the welcome field. And the test program can obtain the current test force of each of the sub-modules according to the module state analysis, and adjust the pressure to be applied by the plurality of sub-modules of the specified hardware module. The test procedure for the stress test method can also be performed inside the module (in other words, the server provided by the present invention can not only perform pressure adjustment between the various hardware modules, but also between the sub-modules 12 201116993). According to the embodiment of the present invention, in the process of performing the pressure job, the hardware module of the feeding device 20 may be unbearably loaded and collapsed or even burned. The pressure testing method of the feeding device may further have a Early warning procedure. Ming Duo..., Fig. 3'' is a hungry diagram of an early warning program according to an embodiment of the present invention. It can be seen from Fig. 3 that the warning program may include step 85: for at least one hard Group definition—pressure warning value; steps such as: reading the module status of the hardware module corresponding to the pressure call alert value; the step says: analyzing the group-like cancer corresponding to the pressure warning value, and obtaining the current pressure warning value Test pressure; step: determine whether the current test pressure corresponding to the pressure alert value is greater than the pressure alert value; and step S54: send a warning signal to the test procedure. In step S50, the warning program may specify a pressure warning value to at least one hardware model, and for example, 5. If the power generated by the power supply 28 is too high, the entire feeding device 20 may be burned, so it should be monitored. . Then, the warning program can read the module state of the hardware module corresponding to the pressure warning value by the BMC 25 in step (6); and analyze the state of the module in step S52 to obtain the corresponding current test pressure. The detailed description of the "module status/current test pressure corresponding to the pressure warning value" refers to the module status/current test pressure _ of the hardware module defined with the pressure warning value. In other words, the early warning program can periodically confirm the module status of the hardware module defining the pressure warning value, and analyze the current test pressure of the hardware module defining the pressure warning value. Then, in step S53', the early warning program determines whether the current test pressure of the hardware module corresponding to the pressure warning value is greater than the pressure warning value. If the current test pressure of this hardware module is greater than the pressure warning, try the program. The early warning procedure is convenient for the step to send the warning signal to the relative test ground, and the test material can be further packaged in the following steps: receiving the hardware model of the current measurement force and the large-scale force of the current force, and in order to avoid the need to control the hardware The module (that is, the hardware model that is defined by the Qianli warning value is faulty, etc., after the 戦 _ _ 录, the force applied to the hardware module. The pre-hoist group can also implement the current scale of the scale In addition, according to the other embodiment of the present invention, the warning program can also set the pressure module to be monitored to better protect the value. 2g. Please refer to Fig. 4'', which is the flow of the warning program according to another embodiment of the present invention, and the figure is shown in Fig. 4." It can be seen that the 'warning procedure may further include the step s6: for at least one The hardware module defines the pressure threshold; the tilting S61: reading the module state of the hardware module corresponding to the critical value of the pressure; step S62: analyzing the module state corresponding to the critical value of the pressure, and obtaining the current test corresponding to the critical value of the pressure Pressure; step S63: judging the corresponding pressure Whether the current test pressure of the threshold is greater than the pressure threshold; and step s64: sending a critical signal to the test program. The difference between the embodiment and the above embodiment is that the warning program further specifies a pressure threshold for the hardware module, and The pressure threshold is higher than the warning value. For example, the central processor's pressure warning value can be set to 80%, and the pressure threshold can be set to 9〇%. The test procedure can further include the following steps: receiving the critical signal And interrupt test procedure. If the current test pressure is higher than the pressure threshold, it indicates that the server 20 has a high probability of system crash or hardware damage in 201116993. Therefore, once the test program receives the critical _, the test procedure Lai temporarily stops the display. After the fresh material is sent = the boundary signal, the data can be backed up to provide the tester's reference. If the pre-supplement program finds the abnormality, the server 2 (4) is closed after the backup data to ensure that the device 20 is Safety. According to the test method of the ship H of the Benedict, the program and the warning program can be executed. The test program and the warning program can be simultaneously used. BMc Μ monitors the module status of the hardware module of the feeding device φ 20 and dynamically adjusts the pressure applied to the hardware module (and its sub-modules) according to the actual situation. The server's test side resources are executed. The test method can also be run on the server management system 3, and the feeder management system 30 can be operated at the far end. In other words, the remote feeder management system can be connected through the network or the Internet. The server 2〇2BMC 25. The pressure test method of the server running on the servo system 30 can also be used to remotely monitor the hardware module of the ship 20 through the BMC 25 and the network remotely, and dynamically adjust for the specified The pressure applied by the hardware module. In summary, the pressure test method of the server according to the embodiment of the present invention may be: the hardware module of the necklace is in a state of being assembled, and the current hardware module is obtained. The pressure is measured and the pressure applied to the specified hardware module is dynamically adjusted to meet the specified test pressure required. In this way, the server's stress test method can adjust the pressure with the actual running condition of the server without interrupting the test program, eliminating many test time and modification scripts due to try an (i error). 201116993 Kung Fu. Further, 'Welcome to the Lai Li Wei Fang to the Qing program to monitor the hardware core group. When the hard test _ line test high force compensation force value or diligent threshold value, the warning material can make the material suitable for t__ Therefore, the early warning procedure can effectively provide the backup test data for the tester to analyze and understand if the accidental situation occurs during the test process. The present invention has been described above with reference to the preferred embodiments thereof, but it is not intended to limit the invention to any skilled artisan, and the present invention may be modified as part of the spirit and scope of the invention. The scope of the remedy shall be determined by the scope of the patent application attached to the specification. [FIG. 1A] FIG. 1A is a server according to an embodiment of the present invention. 1B is a schematic diagram of a server block according to another embodiment of the present invention, and FIG. 2 is a flow chart of a test program according to an embodiment of the present invention. A schematic flowchart of an early warning program according to an embodiment of the present invention; and FIG. 4 is a schematic flow chart of an early warning program according to another embodiment of the present invention. [Description of Main Components] 201116993 20 Server 22 Central Processing Unit 221, 222 Central Processing unit 23 Northbridge wafer 24 Southbridge wafer 25 Baseboard management controller 26 Memory φ 27 Hard disk 28 Power supply 30 Server management system
1717