This page shows you how to manually deploy an insecure multi-node CockroachDB cluster on Google Cloud Platform's Compute Engine (GCE), using Google's TCP Proxy Load Balancing service to distribute client traffic.
The --insecure
flag used in this tutorial is intended for non-production testing only. To run CockroachDB in production, use a secure cluster instead.
To deploy a free CockroachDB Cloud cluster instead of running CockroachDB yourself, see the Quickstart.
Before you begin
Requirements
You must have SSH access to each machine. This is necessary for distributing and starting CockroachDB binaries.
Your network configuration must allow TCP communication on the following ports:
26257
for intra-cluster and client-cluster communication8080
to expose your DB Console
Carefully review the Production Checklist and recommended Topology Patterns.
Do not run multiple node processes on the same VM or machine. This defeats CockroachDB's replication and causes the system to be a single point of failure. Instead, start each node on a separate VM or machine.
To start a node with multiple disks or SSDs, you can use either of these approaches:
- Configure the disks or SSDs as a single RAID volume, then pass the RAID volume to the
--store
flag when starting thecockroach
process on the node. - Provide a separate
--store
flag for each disk when starting thecockroach
process on the node. For more details about stores, see Start a Node.
Warning:If you start a node with multiple--store
flags, it is not possible to scale back down to only using a single store on the node. Instead, you must decommission the node and start a new node with the updated--store
.- Configure the disks or SSDs as a single RAID volume, then pass the RAID volume to the
When starting each node, use the
--locality
flag to describe the node's location, for example,--locality=region=west,zone=us-west-1
. The key-value pairs should be ordered from most to least inclusive, and the keys and order of key-value pairs must be the same on all nodes.When deploying in a single availability zone:
- To be able to tolerate the failure of any 1 node, use at least 3 nodes with the
default
3-way replication factor. In this case, if 1 node fails, each range retains 2 of its 3 replicas, a majority. - To be able to tolerate 2 simultaneous node failures, use at least 5 nodes and increase the
default
replication factor for user data to 5. The replication factor for important internal data is 5 by default, so no adjustments are needed for internal data. In this case, if 2 nodes fail at the same time, each range retains 3 of its 5 replicas, a majority.
- To be able to tolerate the failure of any 1 node, use at least 3 nodes with the
When deploying across multiple availability zones:
- To be able to tolerate the failure of 1 entire AZ in a region, use at least 3 AZs per region and set
--locality
on each node to spread data evenly across regions and AZs. In this case, if 1 AZ goes offline, the 2 remaining AZs retain a majority of replicas. - To ensure that ranges are split evenly across nodes, use the same number of nodes in each AZ. This is to avoid overloading any nodes with excessive resource consumption.
- To be able to tolerate the failure of 1 entire AZ in a region, use at least 3 AZs per region and set
When deploying across multiple regions:
- To be able to tolerate the failure of 1 entire region, use at least 3 regions.
This article covers the use of Linux instances with GCE. You may wish to review the instructions for connecting to Windows instances.
Recommendations
Consider using a secure cluster instead. Using an insecure cluster comes with risks:
- Your cluster is open to any client that can access any node's IP addresses.
- Any user, even
root
, can log in without providing a password. - Any user, connecting as
root
, can read or write any data in your cluster. - There is no network encryption or authentication, and thus no confidentiality.
Decide how you want to access your DB Console:
Access Level Description Partially open Set a firewall rule to allow only specific IP addresses to communicate on port 8080
.Completely open Set a firewall rule to allow all IP addresses to communicate on port 8080
.Completely closed Set a firewall rule to disallow all communication on port 8080
. In this case, a machine with SSH access to a node could use an SSH tunnel to access the DB Console.
Step 1. Configure your network
CockroachDB requires TCP communication on two ports:
26257
for inter-node communication (i.e., working as a cluster)8080
for exposing your DB Console
To expose your DB Console and allow traffic from the TCP proxy load balancer and health checker to your instances, create firewall rules for your project. When creating firewall rules, we recommend using Google Cloud Platform's tag feature to apply the rule only to instances with the same tag.
DB Console
Field | Recommended Value |
---|---|
Name | cockroachadmin |
Source filter | IP ranges |
Source IP ranges | Your local network's IP ranges |
Allowed protocols... | tcp:8080 |
Target tags | cockroachdb |
Application data
Applications will not connect directly to your CockroachDB nodes. Instead, they'll connect to GCE's TCP Proxy Load Balancing service, which automatically routes traffic to the instances that are closest to the user. Because this service is implemented at the edge of the Google Cloud, you'll need to create a firewall rule to allow traffic from the load balancer and health checker to your instances. This is covered in Step 4.
Step 2. Create instances
Create an instance for each node you plan to have in your cluster. If you plan to run a sample workload against the cluster, create a separate instance for that workload.
Run at least 3 nodes to ensure survivability.
Use general-purpose
t2d-standard
,n2-standard
, orn2d-standard
VMs, or use custom VMs. For example, Cockroach Labs has usedt2d-standard-8
,n2-standard-8
, andn2d-standard-8
for performance benchmarking.Warning:Do not use
f1
org1
shared-core machines, which limit the load on CPU resources.If you used a tag for your firewall rules, when you create the instance, click Management, security, disks, networking, sole tenancy. Under the Networking tab, in the Network tags field, enter cockroachdb.
For more details, see Hardware Recommendations and Cluster Topology.
Step 3. Synchronize clocks
CockroachDB requires moderate levels of clock synchronization to preserve data consistency. For this reason, when a node detects that its clock is out of sync with at least half of the other nodes in the cluster by 80% of the maximum offset allowed (500ms by default), it spontaneously shuts down. This avoids the risk of consistency anomalies, but it's best to prevent clocks from drifting too far in the first place by running clock synchronization software on each node.
Compute Engine instances are preconfigured to use NTP, which should keep offsets in the single-digit milliseconds. However, Google can’t predict how external NTP services, such as pool.ntp.org
, will handle the leap second. Therefore, you should:
- Configure each GCE instance to use Google's internal NTP service.
- If you plan to run a hybrid cluster across GCE and other cloud providers or environments, note that all of the nodes must be synced to the same time source, or to different sources that implement leap second smearing in the same way. See the Production Checklist for details.
Step 4. Set up load balancing
Each CockroachDB node is an equally suitable SQL gateway to your cluster, but to ensure client performance and reliability, it's important to use load balancing:
Performance: Load balancers spread client traffic across nodes. This prevents any one node from being overwhelmed by requests and improves overall cluster performance (queries per second).
Reliability: Load balancers decouple client health from the health of a single CockroachDB node. In cases where a node fails, the load balancer redirects client traffic to available nodes.
GCE offers fully-managed TCP Proxy Load Balancing. This service lets you use a single IP address for all users around the world, automatically routing traffic to the instances that are closest to the user.
When using TCP Proxy Load Balancing, you cannot use firewall rules to control access to the load balancer. If you need such control, consider using Network TCP Load Balancing instead, but note that it cannot be used across regions. You might also consider using the HAProxy load balancer (see the On-Premises tutorial for guidance).
To use GCE's TCP Proxy Load Balancing service:
- For each zone in which you're running an instance, create a distinct instance group.
- To ensure that the load balancer knows where to direct traffic, specify a port name mapping, with
tcp26257
as the Port name and26257
as the Port number.
- To ensure that the load balancer knows where to direct traffic, specify a port name mapping, with
- Add the relevant instances to each instance group.
- Configure Proxy Load Balancing.
- During backend configuration, create a health check, setting the Protocol to
HTTP
, the Port to8080
, and the Request path to path/health?ready=1
. This health endpoint ensures that load balancers do not direct traffic to nodes that are live but not ready to receive requests.- If you want to maintain long-lived SQL connections that may be idle for more than tens of seconds, increase the backend timeout setting accordingly.
- During frontend configuration, reserve a static IP address and choose a port. Note this address/port combination, as you'll use it for all of you client connections.
- During backend configuration, create a health check, setting the Protocol to
- Create a firewall rule to allow traffic from the load balancer and health checker to your instances. This is necessary because TCP Proxy Load Balancing is implemented at the edge of the Google Cloud.
- Be sure to set Source IP ranges to
130.211.0.0/22
and35.191.0.0/16
and set Target tags tocockroachdb
(not to the value specified in the linked instructions).
- Be sure to set Source IP ranges to
Step 5. Start nodes
By default, inter-node communication uses the internal IP addresses of your GCE instances.
You can start the nodes manually or automate the process using systemd.
For each initial node of your cluster, complete the following steps:
After completing these steps, nodes will not yet be live. They will complete the startup process and join together to form a cluster as soon as the cluster is initialized in the next step.
Visit Releases and download the full binary of CockroachDB to the node.
On the node, follow the instructions to install CockroachDB.
Run the
cockroach start
command:$ cockroach start \ --insecure \ --advertise-addr=<node1 address> \ --join=<node1 address>,<node2 address>,<node3 address> \ --cache=.25 \ --max-sql-memory=.25 \ --background
This command primes the node to start, using the following flags:
Flag Description --insecure
Indicates that the cluster is insecure, with no network encryption or authentication. --advertise-addr
Specifies the IP address/hostname and port to tell other nodes to use. The port number can be omitted, in which case it defaults to 26257
.
This value must route to an IP address the node is listening on (with--listen-addr
unspecified, the node listens on all IP addresses).
In some networking scenarios, you may need to use--advertise-addr
and/or--listen-addr
differently. For more details, see Networking.--join
Identifies the address of 3-5 of the initial nodes of the cluster. These addresses should match the addresses that the target nodes are advertising. --cache
--max-sql-memory
Increases the node's cache size to 25% of available system memory to improve read performance. The capacity for in-memory SQL processing defaults to 25% of system memory but can be raised, if necessary, to increase the number of simultaneous client connections allowed by the node as well as the node's capacity for in-memory processing of rows when using ORDER BY
,GROUP BY
,DISTINCT
, joins, and window functions. For more details, see Cache and SQL Memory Size.--background
Starts the node in the background so you gain control of the terminal to issue more commands. When deploying across multiple datacenters, or when there is otherwise high latency between nodes, it is recommended to set
--locality
as well. It is also required to use certain enterprise features. For more details, see Locality.For other flags not explicitly set, the command uses default values. For example, the node stores data in
--store=cockroach-data
and binds DB Console HTTP requests to--http-addr=localhost:8080
. To set these options manually, see Start a Node.Repeat these steps for each additional node that you want in your cluster.
For each initial node of your cluster, complete the following steps:
After completing these steps, nodes will not yet be live. They will complete the startup process and join together to form a cluster as soon as the cluster is initialized in the next step.
SSH to the machine where you want the node to run. Ensure you are logged in as the
root
user.Download the CockroachDB archive for Linux, and extract the binary:
$ curl https://binaries.cockroachdb.com/cockroach-v24.2.1.linux-amd64.tgz \ | tar -xz
Copy the binary into the
PATH
:$ cp -i cockroach-v24.2.1.linux-amd64/cockroach /usr/local/bin/
If you get a permissions error, prefix the command with
sudo
.CockroachDB uses custom-built versions of the GEOS libraries. Copy these libraries to the location where CockroachDB expects to find them:
$ mkdir -p /usr/local/lib/cockroach
$ cp -i cockroach-v24.2.1.linux-amd64/lib/libgeos.so /usr/local/lib/cockroach/
$ cp -i cockroach-v24.2.1.linux-amd64/lib/libgeos_c.so /usr/local/lib/cockroach/
If you get a permissions error, prefix the command with
sudo
.Create the Cockroach directory:
$ mkdir /var/lib/cockroach
Create a Unix user named
cockroach
:$ useradd cockroach
Change the ownership of the
cockroach
directory to the usercockroach
:$ chown cockroach /var/lib/cockroach
Download the sample configuration template and save the file in the
/etc/systemd/system/
directory:curl -o insecurecockroachdb.service https://raw.githubusercontent.com/cockroachdb/docs/main/src/current/_includes/v24.2/prod-deployment/insecurecockroachdb.service
Alternatively, you can create the file yourself and copy the script into it:
[Unit] Description=Cockroach Database cluster node Requires=network.target [Service] Type=notify WorkingDirectory=/var/lib/cockroach ExecStart=/usr/local/bin/cockroach start --insecure --advertise-addr=<node1 address> --join=<node1 address>,<node2 address>,<node3 address> --cache=.25 --max-sql-memory=.25 TimeoutStopSec=300 Restart=always RestartSec=10 StandardOutput=syslog StandardError=syslog SyslogIdentifier=cockroach User=cockroach [Install] WantedBy=default.target
Note:Previously, the sample configuration file set
TimeoutStopSec
to 60 seconds. This recommendation has been lengthened to 300 seconds, to give thecockroach
process more time to stop gracefully.In the sample configuration template, specify values for the following flags:
Flag Description --advertise-addr
Specifies the IP address/hostname and port to tell other nodes to use. The port number can be omitted, in which case it defaults to 26257
.
This value must route to an IP address the node is listening on (with--listen-addr
unspecified, the node listens on all IP addresses).
In some networking scenarios, you may need to use--advertise-addr
and/or--listen-addr
differently. For more details, see Networking.--join
Identifies the address of 3-5 of the initial nodes of the cluster. These addresses should match the addresses that the target nodes are advertising. When deploying across multiple datacenters, or when there is otherwise high latency between nodes, it is recommended to set
--locality
as well. It is also required to use certain enterprise features. For more details, see Locality.For other flags not explicitly set, the command uses default values. For example, the node stores data in
--store=cockroach-data
and binds DB Console HTTP requests to--http-port=8080
. To set these options manually, see Start a Node.Start the CockroachDB cluster:
$ systemctl start insecurecockroachdb
Configure
systemd
to start CockroachDB automatically after a reboot:systemctl enable insecurecockroachdb
Repeat these steps for each additional node that you want in your cluster.
systemd
handles node restarts in case of node failure. To stop a node without systemd
restarting it, run systemctl stop insecurecockroachdb
Step 6. Initialize the cluster
On your local machine, complete the node startup process and have them join together as a cluster:
Install CockroachDB on your local machine, if you haven't already.
Run the
cockroach init
command, with the--host
flag set to the address of any node:$ cockroach init --insecure --host=<address of any node on --join list>
Each node then prints helpful details to the standard output, such as the CockroachDB version, the URL for the DB Console, and the SQL URL for clients.
Step 7. Test the cluster
CockroachDB replicates and distributes data behind-the-scenes and uses a Gossip protocol to enable each node to locate data across the cluster. Once a cluster is live, any node can be used as a SQL gateway.
When using a load balancer, you should issue commands directly to the load balancer, which then routes traffic to the nodes.
Use the built-in SQL client locally as follows:
On your local machine, launch the built-in SQL client, with the
--host
flag set to the address of the load balancer:$ cockroach sql --insecure --host=<address of load balancer>
Create an
insecurenodetest
database:> CREATE DATABASE insecurenodetest;
View the cluster's databases, which will include
insecurenodetest
:> SHOW DATABASES;
+--------------------+ | Database | +--------------------+ | crdb_internal | | information_schema | | insecurenodetest | | pg_catalog | | system | +--------------------+ (5 rows)
Use
\q
to exit the SQL shell.
Step 8. Run a sample workload
CockroachDB comes with a number of built-in workloads for simulating client traffic. This step features CockroachDB's version of the TPC-C workload.
Be sure that you have configured your network to allow traffic from the application to the load balancer. In this case, you will run the sample workload on one of your machines. The traffic source should therefore be the internal (private) IP address of that machine.
For comprehensive guidance on benchmarking CockroachDB with TPC-C, see Performance Benchmarking.
SSH to the machine where you want the run the sample TPC-C workload.
This should be a machine that is not running a CockroachDB node.
Download the CockroachDB archive for Linux, and extract the binary:
$ curl https://binaries.cockroachdb.com/cockroach-v24.2.1.linux-amd64.tgz \ | tar -xz
Copy the binary into the
PATH
:$ cp -i cockroach-v24.2.1.linux-amd64/cockroach /usr/local/bin/
If you get a permissions error, prefix the command with
sudo
.Use the
cockroach workload
command to load the initial schema and data, pointing it at the IP address of the load balancer:$ cockroach workload init tpcc \ 'postgresql://root@<IP ADDRESS OF LOAD BALANCER>:26257/tpcc?sslmode=disable'
Use the
cockroach workload
command to run the workload for 10 minutes:$ cockroach workload run tpcc \ --duration=10m \ 'postgresql://root@<IP ADDRESS OF LOAD BALANCER>:26257/tpcc?sslmode=disable'
You'll see per-operation statistics print to standard output every second:
_elapsed___errors__ops/sec(inst)___ops/sec(cum)__p50(ms)__p95(ms)__p99(ms)_pMax(ms) 1s 0 1443.4 1494.8 4.7 9.4 27.3 67.1 transfer 2s 0 1686.5 1590.9 4.7 8.1 15.2 28.3 transfer 3s 0 1735.7 1639.0 4.7 7.3 11.5 28.3 transfer 4s 0 1542.6 1614.9 5.0 8.9 12.1 21.0 transfer 5s 0 1695.9 1631.1 4.7 7.3 11.5 22.0 transfer 6s 0 1569.2 1620.8 5.0 8.4 11.5 15.7 transfer 7s 0 1614.6 1619.9 4.7 8.1 12.1 16.8 transfer 8s 0 1344.4 1585.6 5.8 10.0 15.2 31.5 transfer 9s 0 1351.9 1559.5 5.8 10.0 16.8 54.5 transfer 10s 0 1514.8 1555.0 5.2 8.1 12.1 16.8 transfer ...
After the specified duration (10 minutes in this case), the workload will stop and you'll see totals printed to standard output:
_elapsed___errors_____ops(total)___ops/sec(cum)__avg(ms)__p50(ms)__p95(ms)__p99(ms)_pMax(ms)__result 600.0s 0 823902 1373.2 5.8 5.5 10.0 15.2 209.7
Tip:For more
tpcc
options, usecockroach workload run tpcc --help
. For details about other workloads built into thecockroach
binary, usecockroach workload --help
.To monitor the load generator's progress, open the DB Console by pointing a browser to the address in the
admin
field in the standard output of any node on startup.Since the load generator is pointed at the load balancer, the connections will be evenly distributed across nodes. To verify this, click Metrics on the left, select the SQL dashboard, and then check the SQL Connections graph. You can use the Graph menu to filter the graph for specific nodes.
Step 9. Monitor the cluster
Despite CockroachDB's various built-in safeguards against failure, it is critical to actively monitor the overall health and performance of a cluster running in production and to create alerting rules that promptly send notifications when there are events that require investigation or intervention.
For details about available monitoring options and the most important events and metrics to alert on, see Monitoring and Alerting.
Step 10. Scale the cluster
You can start the nodes manually or automate the process using systemd.
For each additional node you want to add to the cluster, complete the following steps:
SSH to the machine where you want the node to run.
Download the CockroachDB archive for Linux, and extract the binary:
$ curl https://binaries.cockroachdb.com/cockroach-v24.2.1.linux-amd64.tgz \ | tar -xz
Copy the binary into the
PATH
:$ cp -i cockroach-v24.2.1.linux-amd64/cockroach /usr/local/bin/
If you get a permissions error, prefix the command with
sudo
.Run the
cockroach start
command, passing the new node's address as the--advertise-addr
flag and pointing--join
to the three existing nodes (also include--locality
if you set it earlier).$ cockroach start \ --insecure \ --advertise-addr=<node4 address> \ --join=<node1 address>,<node2 address>,<node3 address> \ --cache=.25 \ --max-sql-memory=.25 \ --background
Update your load balancer to recognize the new node.
For each additional node you want to add to the cluster, complete the following steps:
SSH to the machine where you want the node to run. Ensure you are logged in as the
root
user.Download the CockroachDB archive for Linux, and extract the binary:
curl -o cockroach-v24.2.1.linux-amd64.tgz; tar xzvf cockroach-v24.2.1.linux-amd64.tgz
Copy the binary into the
PATH
:cp -i cockroach-v24.2.1.linux-amd64/cockroach /usr/local/bin/
If you get a permissions error, prefix the command with
sudo
.Create the Cockroach directory:
mkdir /var/lib/cockroach
Create a Unix user named
cockroach
:useradd cockroach
Change the ownership of the
cockroach
directory to the usercockroach
:chown cockroach /var/lib/cockroach
Download the sample configuration template:
curl -o insecurecockroachdb.service https://raw.githubusercontent.com/cockroachdb/docs/master/_includes/v24.2/prod-deployment/insecurecockroachdb.service
Alternatively, you can create the file yourself and copy the script into it:
[Unit] Description=Cockroach Database cluster node Requires=network.target [Service] Type=notify WorkingDirectory=/var/lib/cockroach ExecStart=/usr/local/bin/cockroach start --insecure --advertise-addr=<node1 address> --join=<node1 address>,<node2 address>,<node3 address> --cache=.25 --max-sql-memory=.25 TimeoutStopSec=300 Restart=always RestartSec=10 StandardOutput=syslog StandardError=syslog SyslogIdentifier=cockroach User=cockroach [Install] WantedBy=default.target
Note:Previously, the sample configuration file set
TimeoutStopSec
to 60 seconds. This recommendation has been lengthened to 300 seconds, to give thecockroach
process more time to stop gracefully.Save the file in the
/etc/systemd/system/
directoryCustomize the sample configuration template for your deployment:
Specify values for the following flags in the sample configuration template:
Flag Description --advertise-addr
Specifies the IP address/hostname and port to tell other nodes to use. The port number can be omitted, in which case it defaults to 26257
.
This value must route to an IP address the node is listening on (with--listen-addr
unspecified, the node listens on all IP addresses).
In some networking scenarios, you may need to use--advertise-addr
and/or--listen-addr
differently. For more details, see Networking.--join
Identifies the address of 3-5 of the initial nodes of the cluster. These addresses should match the addresses that the target nodes are advertising. Repeat these steps for each additional node that you want in your cluster.
Step 11. Use the cluster
Now that your deployment is working, you can:
- Implement your data model.
- Create users and grant them privileges.
- Connect your application. Be sure to connect your application to the GCE load balancer, not to a CockroachDB node.