Skip to content

Latest commit

 

History

History
32 lines (26 loc) · 2.68 KB

VM_DEPLOY.md

File metadata and controls

32 lines (26 loc) · 2.68 KB
Raw

Deploying Test Machines

This step-by-step instructions below explain how to form a 3-node cluster on Google Cloud Platform, using Virtual Machines (VMs) running Linux. It uses automatic bootstrapping, though there are many other ways to form an rqlite cluster. A similar process can be followed on most Cloud providers. You can also run your rqlite cluster on Kubernetes, and, of course, on bare metal.

Forming the cluster

Start by launching 3 virtual machines, and recording the network (IP) addresses of each.

To automatically bootstrap the rqlite cluster you must know the network (IP) addresses of each VM beforehand. Let's imagine your machines have been assigned network addresses EXTERNAL_IP1, EXTERNAL_IP2, and EXTERNAL_IP3. For this example each node must be reachable from every other node, using the specified network addresses.

Launching rqlite

Next, ssh into each machine and download and install rqlite like so:

export RQLITE_VERSION=v8.32.1
curl -L https://github.com/rqlite/rqlite/releases/download/$RQLITE_VERSION/rqlite-$RQLITE_VERSION-linux-amd64.tar.gz -o rqlite-$RQLITE_VERSION-linux-amd64.tar.gz
tar xvfz rqlite-$RQLITE_VERSION-linux-amd64.tar.gz
sudo cp rqlite-$RQLITE_VERSION-linux-amd64/* /usr/bin

Once installed, run the following command on each node:

VM_IP=`curl -s -H "Metadata-Flavor: Google" http://metadata/computeMetadata/v1/instance/network-interfaces/0/access-configs/0/external-ip`
rqlited -http-addr 0.0.0.0:4001 -http-adv-addr $VM_IP:4001 -raft-addr 0.0.0.0:4002 -raft-adv-addr $VM_IP:4002 -write-queue-batch-size 128 -write-queue-capacity 1024 -bootstrap-expect 3 -join $EXTERNAL_IP1:4002,EXTERNAL_IP2:4002,$EXTERNAL_IP3:4002 data

What does the above command do?

It first gets the IP address of the VM, so that the launch command is identical on each node (this is just for convenience). Then it lauches the rqlite node, telling it to wait for 2 other nodes to contact it before forming a cluster. Each node waits, and whichever node is contacted first by 2 other nodes will form the cluster. Finally, this command also configures the write queue for better indexing performance.

Running the indexer

The indexing program can be run from any VM, or from a 4th VM, or even your local machine. As long as the indexing program can contact one of the nodes in the cluster, it should operate fine.

python indexer.py access.log

By default the indexing program will connect to an rqlite node at localhost:4001. Pass -h to the indexing program to learn how to change this.