cl_status - command for retrieving cluster status information

Cl_status will retrieve many different types of cluster[1] status according to the subcommand given as its first argument.

• hbstatus - indicate if Heartbeat[2] is running on the local system.

• listnodes [-p|-n] - List the nodes[3] in the cluster.

  • -p list only 'ping' type nodes -n list only 'normal' type nodes

• nodestatus <node-name> - list the status of the given node.

• nodetype <node-name> - list the nodes of a given type. Currently the available types are ping and normal.

• listhblinks <node-name> - list the network interfaces used as heartbeat links.

• hblinkstatus <node-name> <link-name> - show the status of a heartbeat link. This indicates up if we are able to hear from that node across that link. Note that if <node-name> is the current node, the status is not meaningful, since with few exceptions we don't receive messages from ourselves on any links.

• clientstatus <node-name> <client-id> [<timeout>] - Show the status of Heartbeat clients. For example, ping or ipfail[4].

• rscstatus - show the status of cluster resource[5]s. Status will be one of: local, foreign, all or none, or transition.

• hbparameter -p <parameter-name> - return value of a given ha.cf[6] parameter.

rscstatus details

Many people find the rscstatus output difficult to understand.

For this example, we'll use the following ActiveActive[7] haresources[8] file

paul 1.2.3.4
silas 5.6.7.8

On paul:

• local == 1.2.3.4
• foreign == 5.6.7.8
• all == 1.2.3.4, 5.6.7.8

On silas:

• foreign == 1.2.3.4
• local == 5.6.7.8
• all == 1.2.3.4, 5.6.7.8

rscstatus relationships

• paul(local) implies silas(local)
• paul(foreign) implies silas(foreign)
• paul(all) implies silas(none) (or silas down)
• paul(none) implies silas(all)
• silas(local) implies paul(local)
• silas(foreign) implies paul(foreign)
• silas(all) implies paul(none) (or paul down)
• silas(none) implies paul(all)

If you issue this command on both nodes (one at a time):    hb_standby local then after this, both machines will be running foreign resources. Weird, but if you want, you can do it. This is about the only way to make a node run only foreign resources.

See Also

mon[9]

---

This information provided courtesy of the Linux-HA project at http://linux-ha.org/