H3C CAS Cloud Management Platform

Configuration Cautions and Guidelines

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Document version: 5W100-20210607

 

Contents

H3C CAS configuration cautions and guidelines· 1

Web-based configuration cautions and guidelines· 1

About Web-based configuration cautions and guidelines· 1

System·· 1

Major: Connecting to the database on CAS· 1

Major: Installing third-party software· 2

Critical: Changing the shared file system heartbeat network· 2

Critical: Mapping a shared storage pool on CAS to a host or VM on another platform. 2

Critical: Reinstalling CAS on a CVK host 2

Major: Upgrading the system·· 3

Host management 3

Critical: Shutting down, powering off, or restarting a host 3

Major: Deleting a host 3

Major: Changing the management IP of a host 3

Major: Editing a vSwitch· 4

Major: Formatting a shared file system·· 4

VM management 4

Major: Powering off a VM·· 4

Minor: Deleting a VM·· 4

Minor: Operating VMs in bulk· 5

Minor: Migrating storage for an online VM·· 5

Major: Repairing storage volumes for a VM·· 5

Major: Migrating storage volumes in bulk· 6

Minor: Edit a VM in a protection group· 6

Minor: Installing a heterogeneous migration client for a VM·· 6

System settings· 6

Major: Editing HA settings· 6

Major: Configuring the action to take on shared storage failure as Do Not Restart Host. 7

Major: Editing NTP settings· 7

CLI-based configuration cautions and guidelines· 8

About CLI-based configuration cautions and guidelines· 8

CLI-based configuration cautions and guidelines· 8

H3C CAS configuration cautions and guidelines

This document contains information that if not or followed can result in severe or undesirable impact on the system or services.
Only trained and qualified personnel are allowed to do the configuration tasks described in this guide.

Before you configure your device, read the information in this document carefully. This document uses the following alert levels to tag information depending on their impact on services:

·     Critical—An alert that calls attention to important information that if not understood or followed can result in system unavailability or service interruption.

·     Major—An alert that calls attention to important information that if not understood or followed can result in unavailability of most functions or services.

·     Minor—An alert that calls attention to important information that if not understood or followed can result in unavailability of some functions without impacting ongoing services.

Web-based configuration cautions and guidelines

About Web-based configuration cautions and guidelines

The following information describes the Web-based tasks that are critical to system functionality and service availability. Only trained and qualified personnel are allowed to do these configuration tasks. If these tasks are not correctly performed, issues such as system shutdown, system reboot, service interruption, service malfunctioning, and deletion of configuration or important files might occur. These issues might cause inaccessibility of the system or services, or cause unexpected logoff of users.

Before you perform a critical task, read the information in this document carefully.

System

Major: Connecting to the database on CAS

Procedure

Log in to the database from the CLI or use third-party database client software to connect to the database.

Impact

A misoperation might cause database corruption and adversely impact the data-related services running in the system.

Guidelines

Before you perform this operation, make sure you are fully aware of impact on the system or contact technical support for help.

Major: Installing third-party software

Procedure

Install non-H3C approved third-party software from the CLI of a CVK host or replace the software provided with the CVK host with non-H3C approved third-party software.

Impact

This operation might cause unpredictable reliability or compatibility issues.

Guidelines

Before you perform this operation, make sure the software is reliable or contact technical support.

Critical: Changing the shared file system heartbeat network

Procedure

Change the shared file system heartbeat network to a non-default network.

Impact

This operation might cause the shared file system to restart because of a fencing event and adversely impact upper layer services.

Guidelines

Before you perform this operation, shut down all VMs in the cluster and suspend the shared file system.

Critical: Mapping a shared storage pool on CAS to a host or VM on another platform.

Procedure

Map a shared storage pool on CAS to a host or VM on another platform.

Impact

This operation will cause storage data conflict and might cause data corruption or loss.

Guidelines

Do not perform this operation.

Critical: Reinstalling CAS on a CVK host

Procedure

Reinstall CAS on a CVK host.

Impact

This operation might adversely impact services running on other hosts in the cluster.

Guidelines

Before you perform this operation, remove the CVK host from the cluster.

Major: Upgrading the system

Procedure

Upgrade the system from a very low version directly to a very high version, for example, upgrade CAS 2.0 or CAS 3.0 directly to CAS 7.0.

Impact

This operation might cause service anomaly after CAS upgrades, if the server is not compatible with the high version of CAS.

Guidelines

Before you perform this operation, verify the software and hardware compatibility of the server with the target version of CAS.

Host management

Critical: Shutting down, powering off, or restarting a host

Procedure

Shut down or restart a host from HDM or by pressing the shutdown or restart button, or power off or remove the power cable from a host while the host is running.

Impact

This operation might adversely impact the services or cause host startup failure.

Guidelines

Before you perform this operation, stop services or suspend the shared storage pool. For more information, contact technical support.

Major: Deleting a host

Procedure

Delete a CVK host from the cluster.

Impact

A misoperation might cause cluster anomaly and adversely impact services running on other hosts in the cluster.

Guidelines

Before you perform this operation, perform the following tasks:

·     Disable cluster HA.

·     Suspend and then delete the shared storage pool.

·     Remove the mapping between the host and the shared storage pool.

Major: Changing the management IP of a host

Procedure

Change the IP address of a host on the management network.

Impact

This operation might cause the host to restart or disconnect from the cluster network because of a fencing event, and eventually adversely impact services.

Guidelines

Before you perform this operation, stop all services on the host, suspend and delete the shared storage pool, and then remove the CVK host from CVM. For more information, contact technical support.

Major: Editing a vSwitch

Procedure

Edit a vSwitch after CVM is available for use.

Impact

This operation might cause network connectivity issues and adversely impact services.

Guidelines

Before you perform this operation, verify the network settings for the vSwitch.

Major: Formatting a shared file system

Procedure

Format a shared file system.

Impact

This operation might cause data loss.

Guidelines

Before you perform this operation, make sure the disk file on the shared file system is not in use.

VM management

Major: Powering off a VM

Procedure

Click Power Off on the VM management page to power off a VM.

Impact

This operation might cause guest OS failure.

Guidelines

Shut down the VM from its guest OS.

Minor: Deleting a VM

Procedure

Delete a VM.

Impact

·     This operation might cause service anomaly.

·     Data files for a VM are permanently deleted on deletion of the VM and they cannot be restored.

·     While the system formats the disk and deletes the data files, storage performance and service might degrade and service issues might occur..

Guidelines

Perform this operation with caution.

·     Before you perform this operation, make sure the VM is not in use, and will not be used any more.

·     Shut down the VM if the VM is running.

·     As a best practice, do not select the formatting the disk and deleting data files option for non-sensitive data. If you have to select this option, make sure the data files will not be used any more.

Minor: Operating VMs in bulk

Procedure

Operate VMs in bulk, for example, starting, restarting, and migrating VMs.

Impact

Bulk VM operations are resources intensive and can degrade host performance.

Guidelines

Perform this operation when no services are running on VMs.

Minor: Migrating storage for an online VM

Procedure

Migrate VM storage while a VM is online.

Impact

·     This operation will rate limit disk I/O and might degrade service performance or cause other service issues.

·     If VM storage migration has not finished before the migration timer expires, the system will suspend the VM. Services running on the VM will become unavailable.

Guidelines

Perform this operation while the service running on the VM is idle.

Major: Repairing storage volumes for a VM

Procedure

Repair the storage volumes and the file system attached to a VM when the VM fails to start because of storage volume errors.

Impact

This operation might cause disk data loss.

Guidelines

Contact technical support to repair the storage volumes.

Major: Migrating storage volumes in bulk

Procedure

Move storage volumes in bulk.

Impact

This operation might cause VM data loss.

Guidelines

Do not move storage volumes in bulk when a VM backup, VM clone, VM snapshot, VM deployment, or DRX task is in progress.

Minor: Edit a VM in a protection group

Procedure

Edit a VM in a protection group.

Impact

This operation will cause VM startup failure if you run a test recovery or failure recovery task.

Guidelines

Do not edit the configuration of the VMs in protection groups in a production environment, unless necessary. Especially, do not perform hardware operations such as adding a vNIC or disk to the VM. If you must edit a VM in a protection group, remove that VM from the group first. After the configuration takes effect, add the VM to the protection group again.

Minor: Installing a heterogeneous migration client for a VM

Procedure

Install a heterogeneous migration client for a VM.

Impact

This operation might cause compatibility issues between the heterogeneous migration client and the CDP backup software from other vendors.

Guidelines

Before you perform this operation, make sure the source VM is not installed with CDP software from a third-party vendor.

System settings

Major: Editing HA settings

Procedure

Edit HA, host heartbeat interval, or storage heartbeat interval settings.

Impact

The stability of CAS might degrade if a too large or too small value is set.

Guidelines

·     Use the default settings for the HA system as long as possible.

·     Do not change the default HA settings for initial deployment.

¡     To change the HA settings for temporary use, evaluate your configuration first and make sure you understand the impact of the change on services.

¡     To change HA settings for long time use, contact technical support to evaluate the impact on services.

Major: Configuring the action to take on shared storage failure as Do Not Restart Host.

Procedure

Configure the action to take on shared storage failure as Do Not Restart Host.

Impact

This operation will trigger a fencing event and cause shared storage volume unmounting error if cluster HA is not enabled.

Guidelines

Enable cluster HA if you configure the action to take on shared storage failure as Do Not Restart Host.

Major: Editing NTP settings

Procedure

Edit NTP settings.

Impact

·     Disabling NTP synchronization might cause the time skew between storage nodes to increase and eventually interrupt services.

·     Changing the NTP server might cause the storage cluster to malfunction if a large time skew exists between the new NTP server and the cluster.

Guidelines

Perform this operation with caution.

·     As a best practice, enable NTP synchronization.

·     Make sure only a small time skew exists between the new NTP server and the cluster.

CLI-based configuration cautions and guidelines

About CLI-based configuration cautions and guidelines

The following information describes the CLI-based tasks that are critical to system functionality and service availability.

Only trained and qualified personnel are allowed to do the configuration tasks described in this guide.

If these tasks are not correctly performed, issues such as system shutdown, system reboot, service interruption, service malfunctioning, and deletion of configuration or important files might occur. These issues might cause inaccessibility of the system or services, or cause unexpected logoff of users.

Before you configure your device, read the information in this document carefully.

CLI-based configuration cautions and guidelines

Operation	Procedure	Impact	Guidelines
System management
Changing the system time of a node	Execute the date -s command to change the system time of a node.	This command might cause cluster errors if the time between nodes in the cluster is out of synchronization.	Critical. Use this command with caution. Make sure you are fully aware of the impact of this command.
Restarting or shutting down a node	Use the following commands to restart, shut down, or power off a node: ·     reboot ·     shutdown ·     poweroff	These commands will cause unavailability of the node and interrupt services running on the node.	Critical. Use these commands with caution. Make sure you are fully aware of the impact of these commands.
Stopping a process	Use the kill or killall command to stop a process.	Incorrect use of this command can cause feature or service anomaly.	Critical. Use these commands with caution. Make sure you are fully aware of the impact of these commands.
Stopping the crond daemon	Use either of the following commands to stop the crond daemon: ·     service crond stop ·     systemctl stop crond.service	The crond daemon executes scheduled commands or scripts in Linux. Stopping the daemon will cause cluster errors.	Critical. Do not execute these commands.
Restarting the libvirt service	Use either of the following commands to restart the libvirt service: ·     service libvirt-bin restart ·     systemctl restart libvirtd.service	These commands will create ghost VMs and damage the disks on the corresponding real VMs if VMs with a Chinese name exist on a CVK host.	Critical. Use these commands with caution. Execute the virsh list --all command at the CLI of the CVK host to verify that no VMs with a Chinese name exist on the host.
Freeing up storage space	Use the fstrim command to free up storage space.	This command will cause a high storage latency, because freeing up storage space requires storage I/O operations.	Major. Use this command with caution. As a best practice, execute this command when no services are running in the system. To obtain help, contact technical support.
Downloading a script	Use the wget command to download a script.	Unpredictable issues might occur if you download an unknown script.	Major. Before executing this command, make sure the script is secure and reliable.
Running an unknown script	None.	Unpredictable issues might occur if you download an unknown script.	Major. Use this command with caution. Before executing this command, verify that the script is secure and reliable and make sure you are fully aware of the impact of the operation.
Changing the system directory or boot file	Use the chroot command to change the system directory or boot file.	This command might cause the host to malfunction.	Critical. Do not execute this command.
Storage management
Running a performance test	Use the fio command to run a storage performance test.	Incorrect use of this command might damage the disks or the shared file system.	Critical. As a best practice, run a performance test before service deployment. If you have to run a performance test after service deployment, contact technical support.
Managing storage	Use the fsmcli command to manage the storage.	Incorrect use of this command might cause service anomaly.	Critical. Do not execute this command.
Disk management
Unmounting a disk	Use the umount command to unmount a disk.	Incorrect use of this command might corrupt the file system and data on a disk and cause service anomaly or data loss.	Critical. Do not execute this command.
Formatting a disk	Use the mkfs command to format a disk.	Incorrect use of this command might cause disk data loss.	Critical. Do not execute this command.
Rebuilding a partition	Use the fdisk or parted command to rebuild a partition.	Incorrect use of this command might clear a partition.	Critical. Use this command with caution. Make sure the host does not have running services before you execute this command. To view all disk partitions, execute the fdisk –l command.
Network configuration
Editing network settings for a node	Use the following commands to edit network settings for a node: ·     ifdown ·     ifup ·     ifconfig down ·     ifconfig up ·     route	These commands will cause network errors on a node and eventually cause cluster errors if your network configuration is incorrect.	Critical. Use these command with caution. Carefully plan your network before executing these commands and make sure you are fully aware of the impact of these commands.
Deleting a vSwitch or vNetwork	Use the following commands to delete a vSwitch or a vNetwork: ·     ovs-vsctl del-br ·     ovs-vsctl del-port	Incorrectly deleting a vSwitch or vNetwork will cause network errors.	Critical. Use the commands with caution. Make sure the target vSwitch or vNetwork can be deleted and delete them from CAS CVM.
Adding a port to a vSwitch	Use the ovs-vsctl add-port command to add a port to a vSwitch.	This command can cause loops if multiple ports are added without being aggregated.	Major. As a best practice, edit vSwitch settings on CAS CVM and verify the network settings for the vSwitch.
Adding or deleting a flow for a vSwitch	Use the following commands to add or delete a flow for a vSwitch: ·     ovs-ofctl add-flow ·     ovs-ofctl del-flows	Incorrect use of this command might cause service interruption and loops.	Major. As a best practice, edit a flow by editing the ACL or vFirewall settings for the VMs on CVM. To edit a flow from the CLI, contact technical support.
VM management
Operating VMs	Use virsh commands to operate VMs.	Use execution commands such as virsh destroy, virsh snapshot, virsh undefine, and virsh migrate with caution. Incorrect use of these commands might adversely impact services. This caution is not true for query commands such as virsh list, virshpool-list, virsh domjobinfo, domblklist, and vncdisplay. If you are not sure whether a command is a query or execution command, do not execute it.	Critical. As a best practice, operate VMs on CVM.
Operating VM disks	Use qemu-img commands to operate VM disks.	Except for query commands such as qemu-img info, execute the qemu-img commands such as qemu-img check, rebase, create, and snapshot with caution. Executing these commands while a VM is running will corrupt its disk files.	Critical. Use these commands with caution. Shut down the VM before you execute these commands.
File management
Changing the group ownership of a file or directory	Use the chgrp command to change the group ownership of a file or directory.	A misoperation might cause cluster errors.	Major. Use this command with caution. Make sure you are fully aware of the impact of this command before executing this command.
Changing the ownership of files or directories	Use the chown command to change the ownership of files or directories.	A misoperation might cause cluster errors.	Major. Use this command with caution. Make sure you are fully aware of the impact of this command before executing this command.
Changing the access permissions of files or directories	Use the chmod command to change the access permissions of files or directories.	A misoperation will cause cluster errors.	Major. Use this command with caution. Make sure you are fully aware of the impact of this command before executing this command.
Removing a directory and its contents recursively	Use the rm –rf command to remove a directory and its contents recursively.	A misoperation might cause important file or data to be mistakenly deleted.	Critical. Use this command with caution. Verify the directory and its contents before executing this command.
Forcibly copying, moving, or creating a file	Use the following commands to copy, move, or create a file: ·     cp –rf ·     scp ·     mv –f ·     touch	Executing these commands will forcibly copy or move a file to a new directory or create a new file in a new directory and overwrite the existing file without any warning or notification.	Major. Use these commands with caution. Make sure the file to copy, move, or create has a different name than the existing files in the new directory.
Operating a file descriptor	Use the echo or dd command to operate a file descriptor.	These commands will edit the files mapped from a disk to the /dev/ directory and might corrupt the file system and data on the disk.	Critical. Do not use these commands.
Editing a file	Use the vi or vim command to edit a file.	A misoperation might cause unpredictable issues.	Major. Use these commands with caution. Make sure you are fully aware of the impact of these commands.
User management
Deleting a user account	Use the userdel command to delete a user account.	Incorrect use of this command might cause service anomaly.	Critical. Do not use this command.
Upgrading a CVK host	Use the ./upgrade.sh –f cvk command to upgrade a CVK host.	This command will remove the CVM component, which is difficult to restore.	Critical. Do not use this command on CVM. To upgrade a CVM host, use the ./upgrade.sh –f cvm command. If the CVM host has been added to the cluster as a CVK host, the CVK host will also be upgraded. If a stateful failover system has been set up, the two CVM hosts will be upgraded.