Configuring LLT manually

Last published : Sep 11, 2026
VCS uses the Low Latency Transport (LLT) protocol for all cluster communications as a high-performance, low-latency replacement for the IP stack. LLT has two major functions.
It handles the following tasks:
  • Traffic distribution
  • Heartbeat traffic
VCS offers support for LLT configuration over bonded NICs, and there are various bonding configurations available for this purpose. During the infoscale configuration, it is necessary to provide the bonded interfaces that are configured on top of private LLT links.
VCS supports the following NIC bonding configurations:
  • Round Robin
  • LACP bonding. However, for LACP bonding, special settings are required at the switch level before configuring NIC bonding using this mode.
When configuring NIC bonding, it is important to note that it is not recommended to provide the individual slave interfaces in the LLT configuration. Instead, it is necessary to provide the bonded interface in the LLT configuration.
VCS also supports bonded NICs over Ethernet and UDP. If high performance is required, NIC bonding is the ideal solution.
Note: During the configuration of LLT (Low Latency Transport), it is important to choose a single type of configuration, either Ethernet or UDP. It is not advisable to mix both of these configurations together.
To configure LLT over Ethernet, perform the following steps on each node in the cluster:
  • Set up the file /etc/llthosts.
See Setting up /etc/llthosts for a manual installation.
  • Set up the file /etc/llttab.
See Setting up /etc/llttab for a manual installation.
  • Edit the following file on each node in the cluster to change the values of the LLT_START and the LLT_STOP environment variables to 1:
/etc/default/llt
You can also configure LLT over UDP.
See Using the UDP layer for LLT.