Skip to main content

Single-node cluster from source

To build from source, clone the repo and run the install script.

$ git clone
$ cd tigerbeetle
$ scripts/

Don't worry, this will only make changes within the tigerbeetle directory. No global changes. The result will place the compiled tigerbeetle binary into the current directory.

Now create the TigerBeetle data file.

$ ./tigerbeetle format --cluster=0 --replica=0 0_0.tigerbeetle
info(io): creating "0_0.tigerbeetle"...
info(io): allocating 660.140625MiB...

And start the server.

$ ./tigerbeetle start --addresses=3000 0_0.tigerbeetle
info(io): opening "0_0.tigerbeetle"...
info(main): 0: cluster=0: listening on

Now you can connect to the running server with any client. For a quick start, try creating accounts and transfers in the Node CLI.

Debugging panics

If TigerBeetle panics and you can reproduce the panic, you can get a better stack trace by switching to a debug build.

$ DEBUG=true scripts/

And then running the server again.