資源資料庫的設計考量Resource pool design considerations

重要

已不再支援此版本的 Operations Manager,建議升級至 Operations Manager 2019This version of Operations Manager has reached the end of support, we recommend you to upgrade to Operations Manager 2019.

資源集區是管理伺服器和/或閘道伺服器的邏輯群組,這些伺服器可用來對彼此發佈工作,並接手失敗成員的工作。A Resource Pool is a logical grouping of management servers and/or gateway servers used to distribute work among themselves and take over work from a failed member. 換句話說,這些伺服器會為工作流程提供高可用性和延展性。In other words, they provide high availability and scalability for workflows. 設計管理群組時,如果要監視網路裝置、Linux/UNIX 系統,以及為充分利用資源集區而設計的其他工作負載,則必須進行一些考量。When designing a management group, considerations must be made for monitoring of network devices, Linux/UNIX systems, and other workloads that are designed to take advantage of a resource pool.

概觀Overview

資源集區提供多個「成員」 來確保持續監視不中斷,這些成員是管理伺服器和 (或) 閘道伺服器,可以在集區的其中一個成員變成無法使用時接手監視工作流程。Resource pools ensure the continuity of monitoring by providing multiple members, which are management servers and/or gateway servers that can take over monitoring workflows if one of the members of the pool becomes unavailable. 您可以建立資源集區以提供特定用途。You can create resource pools for specific purposes. 例如,您可以建立主要資料中心內管理伺服器的資源集區,以監視網路裝置。For example, you might create a resource pool of management servers in your primary data center to monitor network devices.

資源集區會套用類似於叢集「多數節點集」的邏輯,其中 (<當做集區成員的節點數> /2) + 1。Resource pools apply a logic similar to clustering “majority node set”, where (< number of nodes as members of the pool > /2) + 1. 集區中至少必須要有三個成員,才能維持仲裁,即集區中必須有 50% 以上的仲裁投票成員才能維護集區的可用性。At a minimum, there must be three members in the pool to maintain quorum, which must be more than 50% of the quorum voting members in a pool to maintain availability of the pool. 如果您只有集區的兩個成員,但其中一個成員無法使用,則已遺失仲裁。If you only have two members of the pool, and one is unavailable, you have lost quorum.

針對 Operations 主控台中所建立的每個資源集區,Operations Manager 資料庫 (稱為「預設觀察者」 ) 一律會掌握一票 (即使集區中的成員是偶數也是一樣),以便達到仲裁。For every resource pool created in the Operations console, the Operations Manager database, which is referred to as the default observer, is always given a vote, even if you have an even number of members in the pool to allow quorum to be reached. 這也適用於您在初次建立管理群組時預設建立的三個資源集區,將在本主題稍後討論。This also applies to the three resource pools created by default when you first create the management group, which is discussed later in this topic. 針對使用 PowerShell Cmdlet NewSCOM-ResourcePool 所建立的所有資源集區,預設會將它設定為停用。For all resource pools created using the PowerShell cmdlet NewSCOM-ResourcePool, it's set to disabled by default. 包括 Operations Manager 資料庫作為「預設觀察者」 時,您只需要最少部署兩部管理伺服器來維護資源集區的高可用性,即可降低管理群組的複雜度。Including the Operations Manager database as the default observer reduces complexity of your management group by only requiring you to deploy two management servers at a minimum to maintain high availability of your resource pools.

支援資源集區的另一個角色是「觀察者」 。Another role supporting a resource pool are Observers. 這是未參與載入集區之工作流程的管理伺服器或閘道伺服器;不過,它們參與仲裁決策。This is a management server or a Gateway server that doesn't participate in loading workflows for the pool; however they participate in quorum decisions.  這絕不會在正常情況下使用,因此不應該予以考慮。  This is never used under normal circumstances, and therefore shouldn't be considered.

成員資格有兩種類型,分別為自動和手動。There are two types of membership, automatic and manual. 當您建立資源集區時,其成員資格會設定為手動,而且無法重新設定為自動。When you create a resource pool, its membership is set to manual and can't be reconfigured to automatic. 建立 System Center – Operations Manager 管理群組時,預設會建立具有自動成員資格的三個資源集區。When a System Center – Operations Manager management group is created, three resource pools are created by default with automatic membership. 下表說明這三個資源集區。The following table describes these three resource pools.

資源集區名稱Resource Pool Name 說明Description
所有管理伺服器資源集區All Management Servers Resource Pool 執行群組計算、可用性、分散式監視健全狀況彙總,以及資料庫清理的工作流程。Performs workflows for group calculation, availability, distributed monitor health rollup, and database grooming.
通知資源集區Notifications Resource Pool 警示訂閱服務工作流程是以此資源集區為目標,以支援警示通知。The Alert Subscription Service workflows are targeted to this Resource Pool to support alert notifications.
AD 指派資源集區AD Assignment Resource Pool AD 整合工作流程是以此資源集區為目標,以支援對管理伺服器自動指派代理程式。The AD Integration workflows are targeted to this Resource Pool to support automatic agent assignment to management servers.

因為所有管理伺服器資源集區的成員資格都是自動的,所以受委任的任何管理伺服器都會自動成為此資源集區的成員。Because membership of the All Management Servers Resource Pool is automatic, any management server that is commissioned is automatically made a member of this resource pool. 在特定的架構和設計考量 (例如合併地理上分散的應變作業) 中,可能不需要自動指派給所有管理伺服器資源集區。In certain architectures and design considerations, such as those incorporating geographically dispersed contingency operations, automatic assignment to the All Management Servers Resource Pool may not be desired. 在這些情況下,可將成員資格指派從自動變更為手動。In these situations, it's possible to change the membership assignment from automatic to manual. 因此,管理伺服器必須透過手動指派,新增至所有管理伺服器資源集區。As such, management servers must be added to the All Management Servers Resource Pool through manual assignment.

注意

所有管理伺服器資源集區的成員資格均為唯讀。The membership of the All Management Servers Resource Pool is read-only. 若要將其成員資格從自動變更為手動,請參閱修改集區成員資格To change its membership from automatic to manual, see Modifying Pool Membership.

由於導入資源集區之故,建議所有成員都以低延遲網路 (少於 10 毫秒) 連線。With the introduction of resource pools, it is recommended that all members are connected by a low latency network (less than 10 ms). 資源集區不得跨多個資料中心,或在類似 Microsoft Azure 的混合式雲端環境中部署。Resource pools should not be deployed across multiple data centers or in a hybrid-cloud environment like Microsoft Azure.

資源集區可用性範例Resource Pool availability examples

下列範例根據下列僅具有管理伺服器或僅具有閘道伺服器的設定,示範資源集區可用性概念。The following examples demonstrate the concept of resource pool availability based on the following configurations, only with management servers or only with Gateway servers.

單一管理伺服器Single management server

  • 「預設觀察者」 預設會予以啟用,但因只有兩個成員且未達到仲裁而未提供任何優點。The default observer is enabled by default and provides no benefit since there are only two members and quorum isn't reached.
  • 因為管理伺服器是單一失敗點,所以沒有高可用性。There is no high availability, because the management server is a single point of failure.

兩部管理伺服器Two management servers

  • 預設會啟用「預設觀察者」 。The default observer is enabled by default.
  • 因為有三個投票成員,所以集區具有高可用性 - 兩部管理伺服器和「預設觀察者」 。There is high availability for the pool, because there are three voting members - two management servers and the default observer.
  • 如果您停用「預設觀察者」 ,將會失去集區的高可用性。If you disable the default observer, you'll lose high availability for the pool.

三部管理伺服器Three management servers

  • 預設會啟用「預設觀察者」 。The default observer is enabled by default.
  • 因為有四個投票成員,所以集區具有高可用性 - 三部管理伺服器和「預設觀察者」 。There is high availability for the pool, because there are four voting members - three management serves and the default observer.
  • 根據預設,若要維持仲裁,您只能有一部管理伺服器無法使用。By default you can only have one management server unavailable to maintain quorum. 如果兩部管理伺服器無法使用,則您只有 50% 的投票成員,而且資源集區無法再運作來管理監視工作負載。If two management servers are unavailable, you have exactly 50% of voting members and the resource pool no longer functions to manage the monitoring workloads.
  • 「預設觀察者」 不會增加可關閉的管理伺服器數目,因此不會增加集區可用性。The default observer doesn't increase the number of management servers that can be down, therefore it doesn't increase pool availability.
  • 您可以在此情況下考慮移除「預設觀察者」 。You can consider removing the default observer in this scenario.

四部管理伺服器Four management servers

  • 預設會啟用「預設觀察者」 。The default observer is enabled by default.
  • 因為有五個投票成員,所以集區具有高可用性 - 四部管理伺服器和「預設觀察者」 。There is high availability for the pool, because there are five voting members - four management servers and the default observer.
  • 根據預設,若要維持仲裁,您只能有兩部管理伺服器無法使用。By default you can only have two management server unavailable to maintain quorum. 如果三部管理伺服器關閉,則您有 50% 以下的投票成員,而且資源集區無法再運作來管理監視工作負載。If three management servers are down, you have less than 50% of voting members and the resource pool no longer functions to manage the monitoring workloads.
  • 此案例中的「預設觀察者」 提供重要的值,因為它會增加可關閉的管理伺服器數目。The default observer in this scenario provides significant value, because it increases the number of management servers that can be down. 如果沒有「預設觀察者」 ,您只會有四個仲裁成員,而這只允許一個成員無法使用。Without the default observer, you would only have four quorum members, which only allows for one member to be unavailable.

五部管理伺服器Five management servers

  • 預設會啟用「預設觀察者」 。The default observer is enabled by default.
  • 因為有六個投票成員,所以集區具有高可用性 - 五部管理伺服器和「預設觀察者」 。There is high availability for the pool, because there are six voting members - five management servers and the default observer.
  • 根據預設,您只有兩部管理伺服器無法使用,還是可以維持仲裁。By default you can only have two management servers unavailable to maintain quorum. 如果三部管理伺服器無法使用,則這只有 50% 的投票成員,而且資源集區無法再運作來管理監視工作負載。If three management servers are unavailable, this is exactly 50% of voting members, and the resource pool no longer functions to manage the monitoring workloads.
  • 「預設觀察者」 不會增加可關閉的管理伺服器數目,因此不會增加集區可用性。The default observer doesn't increase the number of management servers that can be down, therefore it doesn't increase pool availability.
  • 您可以在此情況下考慮移除「預設觀察者」 。You can consider removing the default observer in this scenario.

在您連繫資源集區中三部以上的管理伺服器之後,即集區中有奇數數目的成員,則可以考慮移除作為成員的「預設觀察者」 。Once you reach three or more management servers in a resource pool, where you have an odd number of members in the pool, you can consider removing the default observer as a member. 如果您聯繫五部管理伺服器,則操作資料庫可能會發生大量負載,而這可能會產生足夠的延遲,足以影響資源集區計算。If you reach five management servers, there is the potential for the Operational database to experience significant load, which might generate enough latency to affect resource pool calculations.

運用「預設觀察者」 扮演角色的方式,集區中的每部管理伺服器都會查詢它自己的本機 SDK 服務,以讓它查詢「預設觀察者」 之操作資料庫中的資料表。With the way the default observer plays a role, each management server in the pool queries its own local SDK service, which allows it to query a table in the Operational database for the default observer. 如果 SDK 服務或資料庫低於負載,將會發生不應該存在的延遲。If the SDK service or database is under a load, you'll experience latency that would otherwise not exist.

單一閘道伺服器Single Gateway server

  • 預設會啟用「預設觀察者」 。The default observer is enabled by default.
  • 因為閘道伺服器是單一失敗點,所以沒有高可用性。There is no high availability because the Gateway server is a single point of failure.
  • 因為閘道伺服器沒有本機 SDK 服務,因此無法查詢操作資料庫,所以不應該在這裡使用「預設觀察者」 。The default observer should not be used here because Gateway servers don't have a local SDK service and therefore can't query the Operational database.

兩部閘道伺服器Two Gateway servers

  • 預設會啟用「預設觀察者」 。The default observer is enabled by default.
  • 因為集區中只有兩個成員,而且「預設觀察者」 因閘道伺服器不會直接與操作資料庫通訊所以參與者,因此沒有高可用性。There is no high availability because there are only two members of the pool and the default observer isn't a participant because Gateway servers don't directly communicate with the Operational database. 需要三部閘道伺服器,才能維持集區仲裁。Three Gateway servers are required to maintain pool quorum.

三部閘道伺服器Three Gateway servers

  • 預設會啟用「預設觀察者」 。The default observer is enabled by default.
  • 因為有三個投票成員,所以集區具有高可用性 - 三部閘道伺服器。There is high availability for the pool, because there are three voting members - three Gateway servers.
  • 根據預設,您只有一部閘道伺服器無法使用,還是可以維持仲裁。By default you can only have one Gateway server unavailable to maintain to maintain quorum. 如果兩部閘道伺服器關閉,則這是 50% 以下的投票成員,而且資源集區無法再運作來管理監視工作負載。If two Gateway servers are down, this is less than 50% of voting members, and the resource pool no longer functions to manage the monitoring workloads.
  • 因為閘道伺服器沒有本機 SDK 服務,因此無法查詢操作資料庫,所以不應該在這裡使用「預設觀察者」 。The default observer should not be used here because Gateway servers don't have a local SDK service and therefore can't query the Operational database.

支援資源集區的監視案例Monitoring scenarios supporting resource pools

下列工作流程是由 Operations Manager 中的資源集區所裝載︰The following workflows are hosted by resource pools in Operations Manager:

  • 網路裝置的管理Management of network devices
  • UNIX/Linux 代理程式的管理Management of UNIX/Linux agents
  • 監視 Web 應用程式 URLMonitoring web application URLs

注意

Windows 代理程式不會回報給資源集區。Windows agents don't report to resource pools.

Operations Manager 中的網路監視需要自己獨立的專用資源集區。Network monitoring in Operations Manager requires its own separate, dedicated resource pool. 這是因為網路監視工作流程是在管理伺服器上 (SNMP 模組上) 執行,而不是在代理程式上執行的緣故。This is because network monitoring workflows run on management servers (on the SNMP module) and not on agents. 一旦您納入網路連接埠的監視 (特別是在您選取裝置上提供的大部分作用中連接埠時) 之後,這將會為管理伺服器帶來沉重的負擔。This places a heavy load on the management servers once you include monitoring of network ports, especially if you select most of the active ports available on the device. 因此,為了提升效能,建議使用專用資源集區中的專用管理伺服器來進行網路監視。Therefore, for better performance, we recommend using dedicated management servers in dedicated resource pools for network monitoring. 此外,屬於此集區成員的管理伺服器應該從所有管理伺服器、通知和 AD 指派集區移除。Additionally, the management servers that are members of this pool should be removed from the All Management Servers, Notifications, and AD Assignment pools.

如果必須啟用高可用性監視和代理程式管理,可以將 Operations Manager 中的 Linux/UNIX 監視指派給專用的資源集區,但這並非必要。Linux/UNIX monitoring in Operations Manager can be assigned to a dedicated resource pool if necessary to enable high-availability monitoring and agent management, but isn't required. Operations Manager 使用憑證來驗證其管理之電腦的存取權。Operations Manager uses certificates to authenticate access to the computers it is managing. 在部署代理程式時,探索精靈會從代理程式擷取、簽署憑證、將憑證重新部署至代理程式,然後重新啟動代理程式。When the Discovery Wizard deploys an agent, it retrieves the certificate from the agent, signs the certificate, deploys the certificate back to the agent, and then restarts the agent. 為支援高可用性,資源集區中的每一部管理伺服器都必須具有用來簽署部署至 UNIX 和 Linux 電腦上代理程式之憑證的所有根憑證。To support high availability, each management server in the resource pool must have all the root certificates that are used to sign the certificates that are deployed to the agents on the UNIX and Linux computers. 否則,若某部管理伺服器變成無法使用,其他管理伺服器將無法信任失敗伺服器所簽署的憑證。Otherwise, if a management server becomes unavailable, the other management servers would not be able to trust the certificates that were signed by the server that failed.

後續步驟Next steps

若要了解如何建立和管理資源集區,請參閱如何管理資源集區To learn how to create and manage resource pools, see How to manage resource pools.