No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>Search

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

FusionInsight HD 6.5.0 Administrator Guide 02

Rate and give feedback :
Huawei uses machine translation combined with human proofreading to translate this document to different languages in order to help you better understand the content of this document. Note: Even the most advanced machine translation cannot match the quality of professional translators. Huawei shall not bear any responsibility for translation accuracy and it is recommended that you refer to the English document (a link for which has been provided).
Creating a Mirror Cluster Using Backup Data

Creating a Mirror Cluster Using Backup Data

Scenario

Create a mirror cluster of the active cluster using backup data when all nodes in the active cluster are offline due to exceptions and cannot provide services, or a test cluster that is completely the same as the active cluster needs to be created.

Create a test cluster that is completely the same as the active cluster in a new network environment.

Prerequisites

  • The network planning of the mirror cluster is the same as that of the active cluster.
  • The number of hosts in the mirror cluster is the same as that in the active cluster. The host names are the same as those in the active cluster.
  • The host disks and partitions in the mirror cluster are the same as those in the active cluster.
  • All IP addresses (including the management IP addresses, service IP addresses, and floating IP addresses of nodes) of the mirror cluster are the same as the corresponding IP addresses of the active cluster.
  • All nodes in the active cluster are offline and cannot provide services if the active cluster and the mirror cluster are connected over a network. Or the mirror cluster is to be installed in an isolated network environment.
  • No patch has been installed in the active cluster. Otherwise, the mirror cluster cannot use backup data to restore properly.
  • The OMS, LdapServer, DBService, Kafka, and NameNode metadata in the active cluster has been backed up to the standby cluster using DistCp, the HBase, Hive, and HDFS service data has been backed up to the standby cluster using DistCp, and the data can be accessed properly. The backup data has been uploaded to and saved in the nodes in the mirror cluster.
NOTE:

The resource pool information displayed on the FusionInsight Manager Tenant Resources page is stored in the HDFS /user/mapred/node-labels directory. You need to back up both HDFS data and data stored in the directory.

Procedure

  1. Select nodes to install the OS of the same version of that on nodes in the active cluster, and run the preinstall script to perform OS preconfiguration and partition mounting.

    For details, see the FusionInsight HD Software Installation. All host names, IP addresses, the number of mounted disks, and disk space must be the same as those of the active cluster.

  2. Install Manager on the management nodes in the mirror cluster.

    On the nodes where the preinstall script is executed, install the active and standby Managers. For details, see Software Installation > Installation Process > Installing Manager on Two Management Nodes. The installation user must be the same as that of the active cluster, such as user root or omm.

    NOTE:
    • The key values in the installation scripts must be the same as those used by the original cluster.
    • If the HA root certificate has been replaced, you need to perform the following operations to add a certificate after the software package is decompressed to the node where the certificate has been replaced (for example, the software package is decompressed to /opt).
      1. Go to the /opt/FusionInsight_Manager/software/hasslCert directory.

        cd /opt/FusionInsight_Manager/software/hasslCert

      2. Prepare a root certificate, and place the new HA root certificate file root-ca.crt and key file root-ca.pem in the /opt/FusionInsight_Manager/software/hasslCert directory.

        The certificate must be the same as the original one.

      3. Set password (empty by default) in the /opt/FusionInsight_Manager/software/hasslCert/hasslCert.ini file to the plaintext password of the root certificate generated.

        For example, set password to Huawei_123. After the cluster is installed, the plaintext password will be deleted.

  3. Log in to the two management nodes as user omm, and run the following command to update the /etc/hosts file on the nodes:

    sh ${BIGDATA_HOME}/om-server/om/sbin/updateOMSHosts.sh

    Information similar to the following is displayed, indicating that the command is executed successfully.

    Succeed to update /etc/hosts.

  4. Log in to the active OMS node as user omm, and run the following commands to use the data backed up before the fault occurs to restore the OMS data (you can view the status of the node by running the $CONTROLLER_HOME/sbin/status-oms.sh script after logging in to the OMS node):

    cd ${CONTROLLER_HOME}/sbin

    sh huaweibigdata_backup.sh -r Full path of OMS backup files

  5. Run the following commands on the active OMS node to disable the active/standby OMS switchover:

    cd ${OMS_RUN_PATH}/workspace/ha/module/hacom/tools/

    ./ha_client_tool --forbidswitch --name=product --time=120

  6. Log in to the active OMS node as user omm, move the LdapServer backup file, for example 6.5.0_LDAP_20180204020014.tar.gz, to the ${BIGDATA_DATA_HOME} directory and ensure that the file user belongs to omm:wheel, and run the following commands to restore the LdapServer data. When the system asks you to enter the password, press Enter.

    cd ${CONTROLLER_HOME}/ldapserver/ldapserver/local/script/

    sh ldapserver_restore.sh -s Full path of LDAP backup files

  7. Log in to the active and standby OMS nodes as user omm, and run the following commands to restart OMS:

    cd ${CONTROLLER_HOME}/sbin

    ./restart-oms.sh

    Run sh ${BIGDATA_HOME}/om-server/om/sbin/status-oms.sh to check whether the value of HAAllResOK of the active management node is Normal. If yes, OMS is restarted successfully.

  8. Reinstall all hosts in the cluster by referring to Reinstalling a Host.

    After Reinstallation, the hosts, topologies, configuration, services, and instances of the mirror cluster are the same as those of the active cluster.

  9. After the cluster hosts is reinstalled, and use the files backed up before the fault occurs to restore the LdapServer data in the cluster. For details, see Recovering Manager Data.
  10. Log in to the active OMS node as user omm, and run the following commands to enable the active/standby OMS switchover:

    cd ${OMS_RUN_PATH}/workspace/ha/module/hacom/tools/

    ./ha_client_tool --cancelforbidswitch --name=product

  11. If a license file has been loaded to the cluster before the fault occurs, log in to FusionInsight Manager and import a license file again.

    For details, see Software Installation > Installation Process > Loading a License File.

  12. Restore the DBService data.

    Restore the DBService data in the cluster. For details, see Recovering DBService Data.

  13. Restore the HBase, HDFS, and Hive data.
Download
Updated: 2019-05-17

Document ID: EDOC1100074522

Views: 6091

Downloads: 12

Average rating:
This Document Applies to these Products
Related Documents
Related Version
Share
Previous Next