Fencing protects your data from being corrupted, and your application from becoming unavailable, due to unintended concurrent access by rogue nodes.
Just because a node is unresponsive doesn’t mean it has stopped accessing your data. The only way to be 100% sure that your data is safe, is to use fencing to ensure that the node is truly offline before allowing the data to be accessed from another node.
Fencing also has a role to play in the event that a clustered service cannot be stopped. In this case, the cluster uses fencing to force the whole node offline, thereby making it safe to start the service elsewhere.
Fencing is also known as STONITH, an acronym for “Shoot The Other Node In The Head”, since the most popular form of fencing is cutting a host’s power.
In order to guarantee the safety of your data [1], fencing is enabled by default.
Note
It is possible to tell the cluster not to use fencing, by setting the stonith-enabled cluster option to false:
[root@pcmk-1 ~]# pcs property set stonith-enabled=false
[root@pcmk-1 ~]# crm_verify -L
However, this is completely inappropriate for a production cluster. It tells the cluster to simply pretend that failed nodes are safely powered off. Some vendors will refuse to support clusters that have fencing disabled. Even disabling it for a test cluster means you won’t be able to test real failure scenarios.
The two broad categories of fence device are power fencing, which cuts off power to the target, and fabric fencing, which cuts off the target’s access to some critical resource, such as a shared disk or access to the local network.
Power fencing devices include:
Fabric fencing devices include:
Using IPMI as a power fencing device may seem like a good choice. However, if the IPMI shares power and/or network access with the host (such as most onboard IPMI controllers), a power or network failure will cause both the host and its fencing device to fail. The cluster will be unable to recover, and must stop all resources to avoid a possible split-brain situation.
Likewise, any device that relies on the machine being active (such as SSH-based “devices” sometimes used during testing) is inappropriate, because fencing will be required when the node is completely unresponsive.
Install the fence agent(s). To see what packages are available, run yum search fence-. Be sure to install the package(s) on all cluster nodes.
Configure the fence device itself to be able to fence your nodes and accept fencing requests. This includes any necessary configuration on the device and on the nodes, and any firewall or SELinux changes needed. Test the communication between the device and your nodes.
Find the name of the correct fence agent: pcs stonith list
Find the parameters associated with the device: pcs stonith describe <AGENT_NAME>
Create a local copy of the CIB: pcs cluster cib stonith_cfg
Create the fencing resource: pcs -f stonith_cfg stonith create <STONITH_ID> <STONITH_DEVICE_TYPE> [STONITH_DEVICE_OPTIONS]
Any flags that do not take arguments, such as --ssl, should be passed as ssl=1.
Enable fencing in the cluster: pcs -f stonith_cfg property set stonith-enabled=true
If the device does not know how to fence nodes based on their cluster node name, you may also need to set the special pcmk_host_map parameter. See man pacemaker-fenced for details.
If the device does not support the list command, you may also need to set the special pcmk_host_list and/or pcmk_host_check parameters. See man pacemaker-fenced for details.
If the device does not expect the victim to be specified with the port parameter, you may also need to set the special pcmk_host_argument parameter. See man pacemaker-fenced for details.
Commit the new configuration: pcs cluster cib-push stonith_cfg
Once the fence device resource is running, test it (you might want to stop the cluster on that machine first): stonith_admin --reboot <NODENAME>
For this example, assume we have a chassis containing four nodes and a separately powered IPMI device active on 10.0.0.1. Following the steps above would go something like this:
Step 1: Install the fence-agents-ipmilan package on both nodes.
Step 2: Configure the IP address, authentication credentials, etc. in the IPMI device itself.
Step 3: Choose the fence_ipmilan STONITH agent.
Step 4: Obtain the agent’s possible parameters:
[root@pcmk-1 ~]# pcs stonith describe fence_ipmilan
fence_ipmilan - Fence agent for IPMI
fence_ipmilan is an I/O Fencing agentwhich can be used with machines controlled by IPMI.This agent calls support software ipmitool (http://ipmitool.sf.net/). WARNING! This fence agent might report success before the node is powered off. You should use -m/method onoff if your fence device works correctly with that option.
Stonith options:
auth: IPMI Lan Auth type.
cipher: Ciphersuite to use (same as ipmitool -C parameter)
hexadecimal_kg: Hexadecimal-encoded Kg key for IPMIv2 authentication
ip: IP address or hostname of fencing device
ipport: TCP/UDP port to use for connection with device
lanplus: Use Lanplus to improve security of connection
method: Method to fence
password: Login password or passphrase
password_script: Script to run to retrieve password
plug: IP address or hostname of fencing device (together with --port-as-ip)
privlvl: Privilege level on IPMI device
target: Bridge IPMI requests to the remote target address
username: Login name
quiet: Disable logging to stderr. Does not affect --verbose or --debug-file or logging to syslog.
verbose: Verbose mode. Multiple -v flags can be stacked on the command line (e.g., -vvv) to increase
verbosity.
verbose_level: Level of debugging detail in output. Defaults to the number of --verbose flags specified
on the command line, or to 1 if verbose=1 in a stonith device configuration (i.e., on
stdin).
debug_file: Write debug information to given file
delay: Wait X seconds before fencing is started
disable_timeout: Disable timeout (true/false) (default: true when run from Pacemaker 2.0+)
ipmitool_path: Path to ipmitool binary
login_timeout: Wait X seconds for cmd prompt after login
port_as_ip: Make "port/plug" to be an alias to IP address
power_timeout: Test X seconds for status change after ON/OFF
power_wait: Wait X seconds after issuing ON/OFF
shell_timeout: Wait X seconds for cmd prompt after issuing command
retry_on: Count of attempts to retry power on
use_sudo: Use sudo (without password) when calling 3rd party software
sudo_path: Path to sudo binary
pcmk_host_map: A mapping of host names to ports numbers for devices that do not support host names. Eg.
node1:1;node2:2,3 would tell the cluster to use port 1 for node1 and ports 2 and 3 for
node2
pcmk_host_list: A list of machines controlled by this device (Optional unless pcmk_host_check=static-
list).
pcmk_host_check: How to determine which machines are controlled by the device. Allowed values: dynamic-
list (query the device via the 'list' command), static-list (check the pcmk_host_list
attribute), status (query the device via the 'status' command), none (assume every
device can fence every machine)
pcmk_delay_max: Enable a random delay for stonith actions and specify the maximum of random delay. This
prevents double fencing when using slow devices such as sbd. Use this to enable a
random delay for stonith actions. The overall delay is derived from this random delay
value adding a static delay so that the sum is kept below the maximum delay.
pcmk_delay_base: Enable a base delay for stonith actions and specify base delay value. This prevents
double fencing when different delays are configured on the nodes. Use this to enable a
static delay for stonith actions. The overall delay is derived from a random delay
value adding this static delay so that the sum is kept below the maximum delay.
pcmk_action_limit: The maximum number of actions can be performed in parallel on this device Cluster
property concurrent-fencing=true needs to be configured first. Then use this to
specify the maximum number of actions can be performed in parallel on this device.
-1 is unlimited.
Default operations:
monitor: interval=60s
Step 5: pcs cluster cib stonith_cfg
Step 6: Here are example parameters for creating our fence device resource:
[root@pcmk-1 ~]# pcs -f stonith_cfg stonith create ipmi-fencing fence_ipmilan \
pcmk_host_list="pcmk-1 pcmk-2" ipaddr=10.0.0.1 login=testuser \
passwd=acd123 op monitor interval=60s
[root@pcmk-1 ~]# pcs -f stonith_cfg stonith
* ipmi-fencing (stonith:fence_ipmilan): Stopped
Steps 7-10: Enable fencing in the cluster:
[root@pcmk-1 ~]# pcs -f stonith_cfg property set stonith-enabled=true
[root@pcmk-1 ~]# pcs -f stonith_cfg property
Cluster Properties:
cluster-infrastructure: corosync
cluster-name: mycluster
dc-version: 2.0.5-4.el8-ba59be7122
have-watchdog: false
stonith-enabled: true
Step 11: pcs cluster cib-push stonith_cfg --config
Step 12: Test:
[root@pcmk-1 ~]# pcs cluster stop pcmk-2
[root@pcmk-1 ~]# stonith_admin --reboot pcmk-2
After a successful test, login to any rebooted nodes, and start the cluster (with pcs cluster start).
| [1] | If the data is corrupt, there is little point in continuing to make it available. |