달력

5

« 2024/5 »

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31

SQL Server 2005 Clustering on Windows Server 2003

1. 클러스터 서비스를 위한 도메인 계정 생성
2. 네트워크 설정
2-1. Public 및 Private IP 주소 할당
2-2. 네트워크 바인딩 순서를 Public을 우선으로 설정
2-3. 인터넷 프로토콜(TCP/IP)를 제외한 모든 구성 요소 체크 해제
2-4. Private 네트워크의 NetBIOS over TCP/IP 기능 해제
2-5. DNS 동적 업데이트 기능 해제
3. 첫 번째 노드에 클러스터 디스크 할당
(ex. 쿼럼(512MB:Q), SQL 데이터(20GB:S), MSDTC(512MB:M))
4. 첫 번째 노드 시스템 종료 후 두 번째 노드에 동일하게 구성
4-1. 드라이브 문자 및 경로 변경
5. 첫 번째 노드 전원 On
6. 각 노드에 DTC, COM+ 설치
(응용 프로그램 서버 > 네트워크 COM+ 액세스 사용, 네트워크 DTC 액세스 사용)
7. 클러스터 생성 (Cluadmin)
7-1. 첫 번째 노드 입력
7-2. 클러스터 IP 입력
7-3. 클러스터 서비스 계정 입력
7-4. 쿼럼 디스크(Q) 설정
7-5. 두 번째 노드 추가
8. 클러스터 네트워크 설정
8-1. Private 네트워크를 '내부 클러스터 통신에만 사용' 설정
8-2. 클러스터 속성 > 네트워크 우선 순위 > Private 우선
9. MSDTC 설정
9-1. MSDTC 그룹 > 새로 만들기 > 리소스 > MSDTC IP 주소, 리소스 종류: IP 주소
9-2. IP 입력, 네트워크: Public
9-3. MSDTC 그룹 > 새로 만들기 > 리소스 > MSDTC 네트워크 이름, 리소스 종류: 네트워크 이름
9-4. MSDTC IP 주소 종속성 추가
9-5. 네트워크 이름 입력, DNS 등록 필수 체크
9-6. MSDTC 그룹 > 새로 만들기 > 리소스 > MSDTC, 리소스 종류: 분산 트랜잭션 코디네이터
9-7. 디스크(M), MSDTC 네트워크 이름 종속성 추가
9-8. 구성 요소 서비스 콘솔 실행(DCOMCNFG)
9-9. 내 컴퓨터 > 속성 > MSDTC > 보안 구성 > 네트워크 DTC 액세스, 원격 클라이언트 허용, 원격 관리 허용, 인바운드 허용, 아웃바운드 허용, 인증 필요 없음, XA 트랜잭션 사용
10. MSDTC 그룹 이동 테스트
11. SQL Server 2005 설치
11-1. SQL Server Database Services, Create a SQL Server failover cluster 체크
11-2. Workstation components, Books Online.... 은 각 노드별로 설치 필요
11-3. 가상 서버 이름 입력,  IP 주소 입력
11-4. SQL 클러스터 그룹 선택
11-5. 두 번째 노드 추가
11-6. 설치 경로 입력, 클러스터 계정 입력, 도메인 그룹 입력
11-7. 설치 진행
12. 서비스팩, 업데이트 롤업 등은 한 노드에만 설치하면 양 노드에 모두 적용됨

:
Posted by 커널64

RMS는 관리 그룹에서 SDK와 Config 서비스를 실행하는 유일한 서버로 이 서비스들이 없이는 관리 그룹이 동작할 수 없다. Operations Manager 데이터베이스를 운영하는 SQL 서버를 클러스터링 하듯이 고가용성을 위해 이 서비스들도 클러스터링이 가능하다.

참고: RMS의 클러스터 노드에는 에이젼트를 설치하지 말아야 한다. 만약 RMS에 대한 모니터링이 필요하다면 다른 관리 그룹에 참가시키거나 에이젼트 없이 관리해야 한다. (Agentless Monitoring)

다음 과정을 통해 관리서버와 관리도구(UI)는 클러스터에 설치되고 Back-end 데이터베이스는 기존에 설치되어 있는(클러스터 되어 있을 수도 있는) SQL 서버를 사용한다.

RMS에 대한 클러스터링 구성 과정은 크게 다음과 같다.
- RMS 클러스터링을 위한 윈도우 클러스터 준비
- RMS 클러스터 그룹에 물리 디스크, IP 주소, 네트워크 이름 리소스를 생성한다.
- 클러스터 노드에 RMS와 두번째 관리서버를 설치하고 RMS 암호키를 백업한다.
- RMS 클러스터 그룹에 RMS의 Health/SDK/Config 서비스에 대해 일반 서비스 리소스를 추가한다.
- Operations Manager 데이터베이스를 백업한다.
- RMS에서 SecureStorageBackup 명령을 실행해 RMS 암호키를 백업한다.
참고: 만약 RMS 설치 완료 시점에서 RMS 암호키가 성공적으로 백업이 됐다면 다시 백업할 필요는 없다.
- 모든 관리서버에서 SecureStorageBackup 명령을 실행해 RMS 암호키를 복원해 기존 RMS 암호키를 설치한다.
- InstallCluster 파라미터를 주고 ManagementServerConfigTool 명령을 실행해 RMS 클러스터 그룹 리소스를 클러스터링한다.
- 모든 RMS 클러스터 그룹을 온라인 시킨다.
- 설정을 완료하기 위해 RMS 클러스터 그룹을 각 노드로 이동시킨다.
- 정상적으로 클러스터 설치가 완료 되었는지 확인한다.
- (옵션) AddRMSNode 파라미터를 주고 ManagementServerConfigTool 명령을 실행해 클러스터 노드를 RMS 클러스터에 추가한다.
- (옵션) 만약 클러스터 설치 과정에서 실패한다면 SetSPN.exe 명령을 실행해야 한다.


클러스터 노드, RMS 클러스터 그룹, RMS 클러스터 그룹 리소스 준비
1. 각 RMS 클러스터 노드에 Operations Manager 관리자 그룹이 로컬 관리자 그룹에 속해있는지, 클러스터의 서비스 계정이 Operations Manager 관리자 그룹의 구성원인지 확인한다.
참고: Operations Manager 관리자 그룹에 클러스터 서비스 계정을 넣는 것은 RMS의 클러스터 구성을 하는데 필요하다.
2. 각 클러스터 노드가 다음과 같은 사전 요구 사항에 맞는지 확인한다.
- Windows Server 2003 SP1 이상
- MDAC 버전 2.80.1022.0 이상
- .NET Framework version 2.0
- .NET Framework version 3.0 구성 요소
3. SDK and Config 서비스 계정을 RMS 클러스터 각 노드의 로컬 관리자 그룹에 추가한다.
4. 최초 RMS의 소유 권한을 가질 노드에 관리자 권한을 가지고 있는 계정으로 로그온한다.
5. 클러스터 관리 도구를 실행(cluadmin)해 클러스터를 생성하고 노드들을 추가한다.
6. 클러스터 관리 도구에서 클러스터 그룹을 새로 만든다. (예: RMSClusterGroup)
7. 기본 소유자 화면에서 RMS 클러스터로 사용할 모든 노드를 추가한다.
8. RMSClusterGroup에 RMS의 IP 주소 리소스를 추가한다. (Public Network, 종속성 없음)
9. '이 주소에 NetBIOS 사용'에 체크가 되어 있는지 확인하고 완료한다.
10. RMSClusterGroup에 실제 디스크 리소스를 추가한다. (종속성 없음)
참고: 모든 클러스터 노드에서 접근이 가능하도록 미리 디스크 파티셔닝 등의 설정이 되어 있어야 한다.
참고: 여기서 사용되는 공유 디스크에는 RMS에 사용되는 MP나 VB 스크립트 등의 데이터가 저장된다.
        Config Service State, Health Service State, SDK Service State 디렉토리가 생성/관리된다.
11. RMSClusterGroup에 네트워크 이름 리소스를 추가한다. (종속성: IP 주소)
12. 매개 변수 페이지에서 RMS의 NetBIOS 이름을 입력하고 'DNS 등록 필수'와 'Kerberos 인증 사용'를 반드시 체크해야 한다.
13. RMSClusterGroup을 온라인 시킨다.


RMS 설치를 위한 Operations Manager의 요구 사항 확인
1. RMSClusterGroup을 소유하고 있는 노드로 로그온한다.
2. Operations Manager의 설치 미디어를 넣고 SetupOM.exe을 실행한다.(설치 초기 화면)
3. 설치 시작 화면에서 '필수 구성 요소 검사'를 클릭해 실행한다.
4. 구성 요소 중 서버와 콘솔을 체크하고 구성 요소 검사를 실행한다.
참고: 이는 Windows Server 2003 SP1, MDAC 버전, .NET Framework 등을 체크한다.
5. 경고는 무시하고 설치할 수 있으나 실패 항목이 있다면 반드시 해결해야 설치가 가능하다.


첫 번째 관리 서버(RMS) 설치
1. 사전에 Back-end 데이터베이스는 미리 설치가 되어 있어야 한다.
2. 설치 구성 요소는 'Management Server'와 'User Interfaces'를 선택한다.
3. SDK 계정을 도메인 계정으로 하는 경우 각 노드의 관리자 그룹에 추가해야 한다.
4. 설치 과정은 일반적인 RMS 설치와 동일하게 설치를 진행한다.
5. 설치 완료 페이지에서 '콘솔 사용' 체크는 제거하고 '암호화 키 백업'은 체크 상태로 둔다.
중요: 설치 과정이 완료되었다고 해서 관리 콘솔을 띄우면 절대로 안 된다.
6. 암호화 키 백업 화면에서 백업을 선택하고 진행한다.
7. 백업 파일의 저장 위치를 지정할 때 반드시 클러스터의 모든 노드에서 접근이 가능한 위치에 저장해야 한다.
8. 암호 입력 후 암호화 키 백업을 완료한다.


두 번째 관리 서버 설치
이 과정에서 나머지 모든 노드에 관리 서버를 설치하게 되며 이 서버들은 과정이 완료될 때까지 이 서버들은 관리 서버이지만 모든 과정이 완료되면 이 서버들은 RMS로 동작할 수 있게 된다.
1. RMS에서 Config 서비스와 SDK 서비스가 실행 중인지 확인하고 Operations Manager 관리자 계정으로 로그온한다.
2. 관리 서버의 설치 과정은 RMS 설치 과정과 동일하며 'Management Server'와 'User Interfaces'를 설치하면 된다.
중요: RMS 설치 때와 마찬가지로 관리 콘솔을 절대로 띄우지 말아야 한다.


RMS 클러스터 리소스 준비
이 과정에서 Operations Manager Health Service (HealthService), Operations Manager Config service (OMCFG), Operations Manager SDK service (OMSDK)에 대한 리소스를 생성한다. 이들 리소스는 IP 주소, 네트워크 이름, 실제 디스크와 함께 클러스터 노드간 장애 복구를 할 수 있게 된다.
1. RMSClusterGroup 클러스터 그룹을 소유하고 있는 노드에 관리 권한을 가진 계정으로 로그온한다.
2. RMSClusterGroup 그룹에 일반 서비스 리소스로 RMS Health Service 리소스를 생성한다. (종속성: RMS 실제 디스크, RMS 네트워크 이름)
3. 매개 변수 페이지에서 서비스 이름을 'HealthService'로 입력한다. 시작 매개 변수는 공란으로 두고 '컴퓨터 이름으로 네트워크 이름 사용'을 체크한다.
4. 레지스트리 복제 페이지의 루트 레지스트리 키 항목은 공란으로 두고 리소스 추가를 완료한다.
5. 나머지 서비스 OMCFG, OMSDK에 대해서도 위와 동일하게 추가한다.
중요: 추가한 서비스 리소스를 온라인 시키면 안 된다.


RMS 클러스터 생성
이 과정에서 나머지 관리 서버에 RMS 키를 배포하고 RMS 클러스터를 생성한다. 이 과정이 완료되면 모든 노드는 RMS를 호스팅할 수 있게 된다.
1. 모든 노드에서 암호화 키의 저장 위치로 접근이 되는지 확인한다.
2. 클러스터 서비스 계정이 Operations Manager Administrators 보안 그룹에 추가되어 있는지 확인한다.
3. 클러스터의 첫 번째 노드인 RMS에 로그온한다.
4. Operations Manager 2007 설치 미디어의 SupportTools 폴더에서 ManagementServerConfigTool.exe 파일을 설치 디렉토리에 복사한다.
5. 위 과정에서 암호화 키의 백업을 하지 않았다면 암호화 키 백업 과정을 진행한다. 저장 위치를 모든 노드가 접속할 수 있는 공유 위치에 저장한다. (예: \\<fileshare>\<filename>.bin)
6. 각 관리 서버에 관리자 계정으로로그온 한다.
7. cmd 창을 열고 설치 디렉토리로 이동 후 다음 명령을 통해 RMS 암호화 키를 복원한다.
SecureStorageBackup.exe Restore \\<fileshare>\<filename>.bin
8. 정상적으로 복원이 되었다면 RMS에 관리자 권한을 가진 계정으로 로그온한다.
9. 클러스터 관리자에서 RMSClusterGroup 클러스터 그룹의 소유자가 RMS인지를 확인한다.
10. Operations Manager 데이터베이스를 호스팅하는 SQL 서버에서 SSMS를 실행해 Operations Manager 데이터베이스의 전체 백업을 수행한다.
중요: RMS 클러스터를 생성하기 위해 ManagementServerConfigTool을 실행하면 Operations Manager 데이터베이스에 복구 불가능한 문제가 발생할 수도 있기 때문이다.
참고: Operations Manager 데이터베이스의 복구 모델은 전체 복구 모델로 설정되어 있어야 한다.
11. RMS에서 cmd 창을 열어 설치 디렉토리로 이동해 다음 명령을 실행한다. 각 클러스터 리소스의 설정에 맞게 입력한다.
ManagementServerConfigTool.exe InstallCluster /vs:<네트워크 이름 리소스의 매개 변수 이름(NetBIOS 이름)> /Disk:<실제 디스크 리소스의 드라이브 명>
12. 복구 불가능한 문제가 발생할 수 있다는 경고 메시지에서 Y를 입력해 진행한다.
13. 'InstallCluster performed successfully' 메시지가 반환되면 정상적으로 완료된 것이다.
참고: ManagmentServerConfigTool.exe InstallCluster 명령은 클러스터의 모든 노드에 클러스터된 RMS를 설치한다.
14. 클러스터 관리자를 열어 RMSClusterGroup을 온라인 시킨다.
15. 다른 노드로 클러스터 그룹을 이동해 정상적으로 온라인되는지 확인한다.


클러스터 설치 확인
1. 관리 콘솔을 열어 관리 페이지로 이동한다.
2. 관리 서버 항목에 RMS의 상태가 정상으로 보이는지 확인한다.
3. 에이젼트 없이 관리 항목에 클러스터의 각 노드의 상태가 정상으로 보이는지 확인한다.

---------------------------------------------------------------------------------------------------------------

클러스터의 RMS 서비스 리소스 이동이 실패하는 관리 서버가 있는 경우
- 해당 관리 서버에 관리자 계정으로 로그온해 서비스 스냅인을 연다.
- OpsMgr SDK Service 속성을 확인해 시작 유형이 '사용 안함'으로 되어 있다면 '수동'으로 변경한다.
- OpsMgr SDK Service를 시작시킨다. (OpsMgr SDK Service는 반드시 RMS에서만 실행되어야 한다.)
- ManagementServerConfigTool.exe InstallCluster /vs:<네트워크 이름 리소스의 매개 변수 이름(NetBIOS 이름)> /Disk:<실제 디스크 리소스의 드라이브 명>


에이젼트 배포 후 모니터링이 되지 않는 경우

증상
Eventlog에 21016, 21023 에러 로그를 남기며 SPN 오류가 나오는 경우
RMS에 대해 SetSPN -L <VirtualManagementServerNetBIOSName> 명령 실행 시
MSomHSvc/<VirtualManagementServerFQDN>가 리스트 되지 않는다.

조치

Support Tools의 SetSPN 명령을 이용해 다음과 같이 RMS의 SPN을 등록한다.
SetSPN.exe -A MSomHSvc/<VirtualManagementServerFQDN> <VirtualManagementServerNetBIOSName>

:
Posted by 커널64

How to Configure/Setup RMS on a Windows 2008 Cluster
1. On each RMS cluster node, ensure that the domain Operations Manager Administrators security group has been added to the local administrators group and that the Cluster service account is a member of the domain Operations Manager Administrators security group.
Note: Having the Cluster service account in the Operations Manager Administrators group is necessary for creating the clustered configuration of RMS.
2. Ensure that each cluster node meets the prerequisites for the Management Server and User Interface components:
- .NET Framework version 3.0 components (Add it from the Features Wizard)
3. Add the SDK and Config service accounts to the Local Administrators group on each node of the RMS cluster.
4. Log on to the cluster node that will be the primary owning node for the RMS with administrative rights.

Cluster preparation
1. In Failover Cluster Management, right click ‘Services and Applications’ and select ‘Configure a Service or Application’.
2. Click next and select ‘Other Server’.
3. Click next and enter the network name for the clustered RMS
4. Click next and select an available storage for cluster group and click next through remainder of wizard.

Install the RMS on the First Node
1. Log on to the cluster node that will be the primary owning node for the RMS with administrative rights
2. Ensure that the RMS cluster group is owned by the node that you are logged onto Proceed with the regular installation wizard
3. On the Completing the System Center Operations Manager 2007 Setup Wizard page, clear the Start the Console checkbox and ensure that the Backup Encryption Key checkbox is selected, then click Finish. The Encryption Key Backup or Restore Wizard will now launch.
Important: Even though the Operations Console has been installed, do not launch the console at this point.
Note If setup fails, it provides you with a value to search on and a link to open the setup log.
4. On the Introduction page of the Encryption Key Backup or Restore Wizard click Next.
5. On the Backup or Restore page select Backup the Encryption Key radio button and click Next.
6. On the Provide a Location page specify a valid path and filename for the encryption key and click Next.
Important: It is critical that the location provided for backing up the encryption key be accessible by all nodes in the cluster.
7. On the Provide a Password page, enter a password to secure the encryption key backup file and click Next to start the backup process.
8. You should now see the Secure Storage Backup Complete page, click Finish.

Install the Secondary Management Servers
In this procedure you will install secondary management servers on all other nodes in the cluster. These servers are secondary Management Servers until this process is complete, at which time they will be able to host the Root Management Server.
1. Log on to each remaining cluster node with the Operations Manager Administrator account.
2. Follow the Install RMS procedures to install the Management Server and User Interface components on each of the other nodes in the Management Group.

Prepare the RMS Cluster Resources
1. In Failover Cluster Management, Right click the RMS server name under ‘Services and Applications’ and select ‘Add a resource -> 4-Generic Service’.
2. From list of services select ‘OpsMgr Health Service’ and click ‘next’ through rest of wizard.
3. Right click newly created resource, choose Properties and enable ‘Use Network Name for computer name’ and add dependencies for the shared disk and network name.
4. Repeat steps 2 and 3 for Config Service and SDK Service.

Create the Virtual RMS
1. Log on to each secondary Management Server computer with an account that is a member of the Administrators group.
2. At a command prompt on each secondary Management Server, type cd <path to Operations Manager installation directory> and then press ENTER.
3. To restore the key to each secondary Management Server, type the following, where <fileshare> is a share accessible by all cluster nodes:
SecureStorageBackup.exe Restore \\<fileshare>\<filename>.bin
Note: You must provide the same password that you entered to encrypt the file on the RMS node.
4. On the SQL server that hosts the OperationsManager database, open the SQL Server Management Studio tool, open the Databases folder and select the OperationsManager database. Right-click to open the context sensitive menu and select Tasks, Back Up to initiate a backup. On the Back Up Database - OperationsManager page, ensure that the Backup type value is set to Full, give the Backup set an appropriate name, and set the Backup set will expire value to a date in the distant future. In the Destination box, for the Back up to value, select Disk and add an appropriate disk location to hold the backup, if one is not already present, and then click OK.
Important: When you run the ManagementServerConfigTool to create the RMS cluster, you are advised to back up the OperationsManager database because irrecoverable damage can be done by creating the RMS cluster.
Note: The OperationsManager database should already be running in the Full Recovery model. For more information, see SQL Server 2005 Books Online.
5. On the RMS server, open a command prompt, type cd <path to Operations Manager installation directory> and then press ENTER.
6. To instantiate the RMS cluster group as a cluster, type the following, where G is the disk resource that is allocated to the cluster group that is being used to create this virtual Root Management Server and where <VirtualServerNetbiosName> is the network name resource allocated to the same cluster group:
ManagementServerConfigTool.exe InstallCluster /vs:<VirtualServerNetbiosName> /Disk:G
The value you enter for <VirtualServerNetbiosName> must be the value that appears in the Name text box located in the Parameters tab of the Properties dialog box for the network name resource.
If InstallCluster runs successfully, the InstallCluster action will return ‘InstallCluster performed successfully’
Note: ManagmentServerConfigTool.exe InstallCluster will install the RMS as a clustered service on every available node in the cluster.
7. In the Cluster Administrator, right click the RMSClusterGroup to open the context menu and select Bring Online. This will bring all the RMSClusterGroup services online.

:
Posted by 커널64

-전체적인 순서-

1. 모든 리소스는 Active Node에 있어야 한다. (Quorum, Exchange)
2. Passive Node를 먼저 SP1으로 Upgrade 한다.
3. CMS 리소스를 중지하고 CMS 리소스를 Passive Node로 이동한 후 CMS를 Upgrade한다. CMS 리소스를 Online 시킨다.
4. Active Node였던 현재의 Passive Node를 Upgrade한다.
 

-상세 과정-

1. 모든 리소스는 Active Node에 있어야 한다. (Quorum, Exchange)
1-1. Move-ClusteredMailboxServer cmdlet을 이용해 Upgrade할 서버에서 다른 Node로 CMS를 이동한다.
1-2. 만약 Upgrade할 서버에 클러스터 그룹 리소스(Quorum)이 있다면 클러스터 관리자를 이용해 다른 Node로 이동한다.

2. Passive Node를 먼저 SP1으로 Upgrade 한다.
2-1. 다음 작업은 Passive Node에서 이루어진다.
2-2. 성능 카운터를 Handling하는 Service 또는 Process들을 중지시킨다. (예: MOM Agent)
2-3. Powershell에서 다음 명령을 이용해 Windows Firewall을 시작하고 Remote Registry 서비스를 재시작한다.
Set-Service SharedAccess –StartupType:Manual
Start-Service SharedAccess
Restart-Service RemoteRegistry
2-4. 설치를 시작하기 전에 Exchange Server와 관련된 Process를 종료한다. (예:Management Console, Management Shell)
2-5. Exchange Server 2007 SP1 설치 미디어를 넣고 다음 명령을 이용해 Passive Node를 Upgrade한다.
setup.com /mode:upgrade
2-6. 설치 완료 후 재부팅한다.

3. CMS 리소스를 Offline 시키고 CMS 리소스를 Passive Node로 이동한 후 CMS를 Upgrade한다.
3-1. 재부팅이 되면 Exchange Management Shell에서 다음 명령을 이용해 CMS를 중지한다.
Stop-ClusteredMailboxServer <CMS Name> -StopReason "Upgrade to SP1"
3-2. 클러스터 관리도구를 이용하여 모든 리소스를 Upgrade한 Passive Node로 이동한다.
3-2. 이제 Passive Node는 SP1으로 Upgrade되었으며 CMS 리소스를 소유하고 있는 상태이다.
3-3. 다음 명령을 이용해 CMS를 SP1으로 Upgrade한다. (CMS의 Upgrade가 완료되면 자동으로 리소스가 Online 된다.)
setup.com /UpgradeCMS
3-4. 기존에 Active Node였던 Passive Node에 대하여 2-2부터 2-6과정을 진행한다.


위 과정은 CCR의 과정이나 SCC의 경우도 동일한 방식으로 진행하면 된다.

마찬가지로 Passive Node를 먼저 Upgrade시키고 CMS 리소스를 이동 후 Upgrade하고 나머지 Node들을 Upgrade한다.

:
Posted by 커널64

Quorum 파일 구성

<Quorum Disk>:\MSCS\Chkxxx.tmp

<Quorum Disk>:\MSCS\Quolog.log

<Quorum Disk>:\MSCS\<Resource GUDI>\*.CPT

 

Chkxxx.tmp, Quolog.log 손상

- ClusSvc Event ID 1148 and 1073 : Chkxxx.tmp and Quolog.log is corrupt

 

[Resolution]

- System State backup을 복구

- Cluster 서비스 시작 옵션 이용

 

net start clussvc /NoQuorumLogging

-> 서비스가 시작되면 아래의 절차 수행

-> 서비스가 시작되지 않는다면 Cluster 디스크 드라이버, Cluster 네트웍 드라이버를 Disable 시킨 후

   서버를 Reboot하여 Quorum Disk가 물리적으로 정상인지 확인

net stop clussvc

net start clussvc /ResetQuorumLog -> 현재 쿼럼 로그를 초기화하고 새로 작성

net stop clussvc

net start clussvc

 

 

클러스터 서비스 시작 옵션

net start clussvc /FixQuorum

Quorum 디스크의 물리적인 문제가 발생했을 경우 최소 1노드에서 서비스 가능케 함

 

이 경우,

Quorum 디스크를 제외한 다른 클러스터 디스크는 정상 동작해야 함

Cluster 리소스 그룹은 온라인 되지 않음

다른 노드는 Cluster에 조인하지 못함

Cluster 관리자를 아래 2가지 방법으로 실행

Cluadmin -noreconnect or . (localhost)

:
Posted by 커널64

Cluster DB 파일 손상

- System Event Log : ClusSvc Event ID 1000

- Cluster Log : Failed to open key resources, status 2

 

[Resolution]

Case1 - 클러스터 서비스가 최소 다른 한 노드에서 시작 중인 경우

1) 정상 노드에서 Q:\MSCS에 접근하여 Chkxxx.tmp Copy한다.

2) Copy한 파일을 문제가 발생한 노드의 %SystemRoot%\Cluster Paste한다.

3) 문제의 ClusDB 파일을 Rename 한다.

4) 새로 Copy Chkxxx.tmp ClusDB Rename한다.

5) Cluster 서비스를 시작한다.

 

Case2 - 클러스터 서비스가 어떠한 노드에서도 시작 중이지 않은 경우

1) 한 노드에서 Cluster 디스크 드라이버를 중지시킨다. (장치 관리자 -> 숨김 장치)

2) 모든 노드를 shutdown 한다.

3) Cluster 디스크 드라이버를 중지시킨 노드를 Turn on하여 Chkxxx.tmp Q:\MSCS에서 Copy한다.

4) Chkxxx.tmp %SystemRoot%\Cluster Paste한다.

5) 기존 ClusDB 파일은 rename하거나 Delete한다.

6) Chkxxx.tmp 파일을 ClusDB rename한다.

7) 클러스터 서비스를 시작한다.

8) 클러스터 서비스가 시작되면, Case1을 반복 진행한다.

:
Posted by 커널64