Generally, connection issues fall into one of the following three areas:
- Connecting - are you able to reach your instance over the network?
- Authorizing - are you authorized to connect to the instance?
- Authenticating - does the database accept your database credentials?
Each of those can be further broken down into different paths for investigation. The following section includes examples of questions you can ask yourself to help further narrow down the issue:
Connection issues checklist
- Private IP
- Have you enabled the
Service Networking APIfor your project?
- Are you using a Shared VPC?
- Does your user or service account have the required IAM permissions to manage a private services access connection?
- Is private services access connection configured for your project?
- Did you allocate an IP address range for the private connection?
- Is the private connection created?
- If the private connection was changed, were the vpc-peerings updated?
- Do the VPC logs indicate any errors?
- Is your source machine's IP a non-RFC 1918 address?
- Can you traceroute from end to end, from source to destination?
- Public IP
- Cloud SQL Auth proxy
- Is the Cloud SQL Auth proxy up to date?
- Is the Cloud SQL Auth proxy running?
- Is the instance connection name formed correctly in the Cloud SQL Auth proxy connection command?
- Have you checked the Cloud SQL Auth proxy output? Pipe the output to a file, or watch the terminal where you started the Cloud SQL Auth proxy.
- Does your user or service account have the required IAM permissions to connect to a Cloud SQL instance?
- Have you enabled the
Cloud SQL Admin APIfor your project?
- If you have an outbound firewall policy, make sure it allows connections to port 3307 on the target Cloud SQL instance.
- If you are connecting using UNIX domain sockets, confirm that the sockets were created by listing the directory specified with the -dir when you started the Cloud SQL Auth proxy.
- Cloud SQL connectors and language-specific code
- Is the connection string formed correctly?
- Have you compared your code with the sample code for your programming language?
- Are you using a runtime or framework for which we don't have sample code?
- If so, have you looked to the community for relevant reference material?
- Self-managed SSL/TLS certificates
- Is the server certificate still valid?
- Authorized networks
- Is the source IP address included?
- Are you using a non-RFC 1918 IP address?
- Are you using an unsupported IP address?
- Connection failures
For specific API error messages, see the Error messages reference page.
Additional connectivity troubleshooting
For other issues, see the Connectivity section in the troubleshooting page.
Common connection issues
Verify that your application is closing connections properly
If you see errors containing "
Aborted connection nnnn to db:", it usually
indicates that your application is not stopping connections properly. Network issues can also cause this error. The error does not mean that
there are problems with your Cloud SQL instance. You are also encouraged to run
tcpdump to inspect the packets to track down the source of the problem.
For examples of best practices for connection management, see Managing database connections.
Verify that your certificates have not expired
If your instance is configured to use SSL, go to the Cloud SQL Instances page in the Cloud Console and open the instance. Open its Connections page and make sure that your server certificate is valid. If it has expired, you must add a new certificate and rotate to it.
Verify that you are authorized to connect
If your connections are failing, check that you are authorized to connect:
- If you are having trouble connecting using an IP address, for example,
you are connecting from your on-premises environment with the sqlcmd
client, then make sure that the IP address you are connecting from
is authorized to connect
to the Cloud SQL instance.
Connections to a Cloud SQL instance using a private IP address are automatically authorized for RFC 1918 address ranges. This way, all private clients can access the database without going through the Cloud SQL Auth proxy. Non-RFC 1918 address ranges must be configured as authorized networks.
Cloud SQL doesn't learn Non-RFC 1918 subnet routes from your VPC by default. You need to update the network peering to Cloud SQL to export any Non-RFC 1918 routes. For example:
gcloud compute networks peerings update cloudsql-mysql-googleapis-com \ --network=NETWORK \ --export-subnet-routes-with-public-ip \ --project=PROJECT_ID
Here's your current IP address.
Determining how connections are being initiated
You can see information about your current connections by connecting to your database and running the following command:
Connections that show an IP address, such as
126.96.36.199, are connecting using IP.
cloudsqlproxy~188.8.131.52 are using the Cloud SQL Auth proxy, or else they
originated from App Engine. Connections from
localhost may be
used by some internal Cloud SQL processes.
Understanding connection limits
There are no QPS limits for Cloud SQL instances. However, there are connection, size, and App Engine specific limits in place. See Quotas and Limits.
Database connections consume resources on the server and the connecting application. Always use good connection management practices to minimize your application's footprint and reduce the likelihood of exceeding Cloud SQL connection limits. For more information, see Managing database connections.
Show connections and threadsTo see the processes that are running on your database, connect to your database and run the following command:
For information about how to interpret the columns returned from
sp_who, see the SQL Server reference.
Connections timeout (from Compute Engine)
Connections with a Compute Engine instance timeout after 10 minutes of inactivity, which can affect long-lived unused connections between your Compute Engine instance and your Cloud SQL instance. For more information, see Networking and Firewalls in the Compute Engine documentation.
To keep long-lived unused connections alive, you can set the TCP keepalive. The following commands set the TCP keepalive value to one minute and make the configuration permanent across instance reboots.
Display the current tcp_keepalive_time value.
Set tcp_keepalive_time to 60 seconds and make it permanent across reboots.
echo 'net.ipv4.tcp_keepalive_time = 60' | sudo tee -a /etc/sysctl.conf
Apply the change.
sudo /sbin/sysctl --load=/etc/sysctl.conf
Display the tcp_keepalive_time value to verify the change was applied.
Tools for debugging connectivity
The most basic tools are
Ping performs a basic test to determine if the destination
("Cloud SQL instance") is available from the source.
Ping sends an
ICMP Echo Request packet to a Cloud SQL instance, and it expects an
ICMP Echo Reply in return. If
ping doesn't succeed,
then there's no route from the source to the destination. Success, however,
doesn't mean that your packets can get through, only that in general, the
Cloud SQL instance can be reached.
ping can tell if a host is alive and responding, it's not
guaranteed to be reliable. Some network providers block
ICMP as a
security precaution, which can make connectivity debugging more difficult.
Traceroute tests the complete route network packets take from one host
to another. It shows all the steps ("hops") that the packet takes
along the way, and how long each step takes. If the packet doesn't go through
all the way to the destination,
traceroute doesn't complete, but
ends with a series of asterisks. In this case, look for the last IP address that
was successfully reached along the way. This is where connectivity broke down.
Traceroute can time out. It can also fail to complete if a gateway
along the way isn't configured correctly to pass the packet along to the next
traceroute fails to complete, you might be able to figure out
where it stopped. Find the last IP address listed in the
output, and do a browser search for
who owns [IP_ADDRESS]. Results
may or may not show the owner of the address, but it's worth a try.
mtr tool is a form of
traceroute that remains
live and continuously updated, similar to how the
works for local processes.
tcpdump is a tool to capture packets. It is highly encouraged to run
tcpdump to capture and inspect the packets between your host and the CloudSQL instances when you are debugging the connectivity problems.
Locate your local IP address
If you don't know the local address of your host, then run the
ip -br address show command. On Linux, this shows the network interface,
the status of the interface, the local IP, and MAC addresses. For example:
eth0 UP 10.128.0.7/32 fe80::4001:aff:fe80:7/64.
Alternatively, you can run
ifconfig to see
the status of your network interfaces.
Testing with Connectivity Test
Connectivity Test is a diagnostics tool that lets you check connectivity between endpoints in your network. It analyzes your configuration and in some cases performs run-time verification. It supports CloudSQL as a preview feature. Follow these instructions to run tests with your CloudSQL instances.
Testing your connection
You can use the sqlcmd client to test your ability to connect from your local environment. For more information, see Connecting the sqlcmd client using IP addresses and Connecting the sqlcmd client using the Cloud SQL Auth proxy.
Determining the IP address for your application
To determine the IP address of a computer running your application so you can authorize access to your Cloud SQL instance from that address, use one of the following options:
- If the computer is not behind a proxy or firewall, log in to the computer and use this link to determine its IP address.
- If the computer is behind a proxy or firewall, log in to the computer and use a tool or service like whatismyipaddress.com to determine its true IP address.
Open local ports
To verify that your host is listening on the ports you think it is, run the
ss -tunlp4 command. This tells you what ports are open and
All local port activity
netstat command to see all the local port activity. For
netstat -lt shows all the currently active ports.
Connect to your Cloud SQL instance using telnet
To verify that you can connect to your Cloud SQL instance using
telnet command. Telnet attempts to connect to the IP address and
port you give it.
On success, you see the following:
Connected to 184.108.40.206.
Connected to 220.127.116.11.
On failure, you see Trying 18.104.22.168...
telnet hangs until you force-close the attempt:
You can view logs for Cloud SQL instances and other Google Cloud projects such as Cloud VPN or Compute Engine instances. To view logs for your Cloud SQL instance log entries:
In the Google Cloud Console, go to the Cloud Logging page.
- Select an existing Cloud SQL project at the top of the page.
- In the Query builder, add the following:
- Resource: Select Cloud SQL Database. In the dialog, select a Cloud SQL instance.
- Log names: Scroll to the Cloud SQL section and select
appropriate log files for your instance. For example:
- Severity: Select a log level.
- Time range: Select a preset or create a custom range.
Private IP addresses
Connections to a Cloud SQL instance using a private IP address are automatically authorized for RFC 1918 address ranges. Non-RFC 1918 address ranges must be configured in Cloud SQL as authorized networks. You also need to update the network peering to Cloud SQL to export any Non-RFC 1918 routes. For example:
gcloud compute networks peerings update cloudsql-sqlserver-googleapis-com
The IP range 172.17.0.0/16 is reserved for the Docker bridge network. Any Cloud SQL instances created with an IP address in that range will be unreachable. Connections from any IP address within that range to Cloud SQL instances using private IP address will fail.
See the Cloud VPN troubleshooting page.