LLT (Low latency Transport) should be startup first using the “lltstat -c” command. It reads /etc/llttab and /etc/llthosts files and establishes heartbeat network. Heartbeat network is a private network where VCS status information is exchanged by all systems within a VCS cluster. These networks require each system in the cluster to have a dedicated NIC, connected to a private hub. VCS requires a minimum of two dedicated communication channels between each system in a cluster. LLT is a low overhead networking protocol that runs in the kernel. Because it runs in the kernel, it is capable of handling kernel to kernel communications.
Examples of files are as below:
#cat /etc/llthosts 0 <node 1> 1 <node 2> . . n <node n+1>In the Example below, the linux systems will have interface names such as “eth0/1″. If using any device, then replace “ce” with “qfe0/1″ etc..
#cat /etc/llttab set-node set-cluster link ce2 /dev/ce:0 - ether - - link ce3 /dev/ce:3 - ether - - link-lowpri ce4 /dev/ce:4 - ether - - start
Verification for startup can be done using lltstat -n command. “*” represents firstnode (master) in the cluster
#lltstat -n Node State Links * 0 <node1> OPEN 3 1 <node2> OPEN 3 2 <node3> OPEN 3 . . n <node n+1> OPEN 3
GAB (Group Atomic Board) starts next. It executes /etc/gabtab and checks for other GABs to establish a cluster membership. GAB runs over Low Latency Transport (LLT) and uses broadcasts to distribute cluster configuration information and ensure that each system has a synchronized view of the cluster, including the state of each system, service group, and resource.
# cat /etc/gabtab /sbin/gabconfig -c -n5 # for a 5 node cluster
GAB can be started using “gabconfig -c” and verified by using “gabconfig -a”. Below is the example output for a 5 node cluster. Port ‘a’ runs GAB service and port ‘h’ runs VCS Deamon
# gabconfig -a GAB Port Memberships ========================================= Port a gen 11ff05 membership 012345 Port h gen 11ff09 membership 012345
After both LLT and GAB are loaded, hashadow starts which will load HAD (High Availability Deamon). HAD reads /etc/VRTSvcs/conf/config/main.cf, types.cf and all include.cf’s mentioned in main.cf file.
HAD checks if there are other HADs avaible and registers them with GAB. If there are no other HADs, it loads the main.cf again into HAD memory.Same process will happen when HAD starts on
other nodes. The HAD on the first node will load the main.cf and other include.cf files from the local system and all other HADs will load configuration from the first HAD.
After starting up, HAD will know all the service groups and resources from main.cf. It will call the respective agents to check if the resources are currently online or offline. Based on
main.cf, HAD will online/offline the Service group on the respective nodes.
Cluster is started up with “hastart” command. The status can be verified using “hastatus -sum”
VCS Logfile: /var/<vcs-path>/log/engine_A.log