Upgrade Kernel

Procedure

1. Update the local repo of the management node(s).
   Make sure that you have update the local repo of the management node(s) with the new ISO. If not, run the following command on the management node(s). If you deploy dual management nodes, run the command on the two management nodes respectively:

   [root@localhost ~]# zstack-upgrade -r ZStack-Cloud-x86_64-DVD-5.3.20-h84r.iso

2. Back up the database of the management node(s).
   Make sure that you have backed up the database of the management node(s). If not, run the following command on the management node(s) to backup the database to /var/lib/zstack/mysql-backup/lib/zstack/mysql-backup/. If you deploy dual management nodes, run the command on the two management nodes respectively:

   [root@localhost ~]# zstack-ctl dump_mysql --file-name zstack-db-backup

3. Check ZStack Cloud services and hardware connectivity.
   Make sure that the ZStack Cloud service is normal and all hardware devices are connected well.
4. Disable the VM high availability globally.
   Log in to the ZStack Cloud UI. On the main menu, choose Setting > Platform Setting > HA Policy. On the HA Policy page, click the Disable button on the upper right corner.
5. Enable the global setting zstack.experimental.repo.
   Run the following commands on the management node to enable the global setting zstack.experimental.repo. If you deploy dual management nodes, run on the management node that corresponds to the VIP:

   [root@localhost ~]# zstack-cli # Lanuch the CLI command tool - >>>  LogInByAccount accountName=admin password=${Password} # Log in the admin account. admin >>> UpdateGlobalConfig category=cluster name=zstack.experimental.repo value=true # Enable the global setting zstack.experimental.repo.

   Note: If you deploy dual management nodes, you can run zsha2 status on either of the management nodes. The management node whose VIP value is yes is the management node you need.
6. Upgrade host Kernel.
   Run the following CLI commands on the management node to upgrade host Kernel:

   admin >>> QueryCluster    //Query UUIDs of clusters. admin >>> UpdateClusterOS updatePackages=kernel excludePackages=python2-crypto uuid=${ClusterUuid}    //Upgrade Kernel of hosts in the specified cluster. Record the uuid (LongJobUuid) returned. admin >>> QueryLongJob uuid=${LongJobUuid}    //Query the upgrading progress. If the returned state value is Succeeded, the upgrading is completed.

   Note: If you deploy more than one cluster, you can upgrade the Kernel of host in different clusters sequentially.
7. Check the Kernel version after the upgrading.
   Run the following command on the hosts to check whether the Kernel are upgraded to the target version:

   [root@localhost ~]# rpm -qa | grep -i kernel

8. Reboot hosts.
   Reboot hosts to make the new Kernel version take effect.

   Note: Rebooting a host temperately interrupts VM business running on it. We recommend that you perform this action during off-peak periods.
9. Disable the global setting zstack.experimental.repo.
   After the upgrading, run the following CLI command on the management node to disable the global setting zstack.experimental.repo:

   admin >>> UpdateGlobalConfig category=cluster name=zstack.experimental.repo value=false

10. Enable the VM high availability globally.
    On the main menu of ZStack Cloud, choose Setting > Platform Setting > HA Policy. On the HA Policy page, click the Enable button on the upper right corner.