Diagnosing Issues with Cloud SQL Instances

This page contains a list of the most frequent issues you might run into when working with Google Cloud SQL instances as well as steps you can take to address them. If the information here does not solve your issue, see the Support Overview for getting further help.

Viewing logs

To see information about recent operations, you can view the Cloud SQL instance operation logs or the MySQL error logs.

Connection issues

Verify that you are authorized to connect

If your connections are failing, check that you are authorized to connect:

  • If you are connecting from App Engine standard environment to a First Generation instance, ensure that the App Engine application is authorized to connect to the Cloud SQL instance.
  • If you are having trouble connecting using an IP address, for example, you are connecting from your on-premises environment with the mysql client, then make sure that the IP address you are connecting from is authorized to connect to the Cloud SQL instance. Here's your current IP address.
  • Try the gcloud sql connect to connect to your instance. This command authorizes your IP address for a short period of time. You can run this in an environment with Cloud SDK and mysql client installed. You can also run this command in Google Cloud Shell, which is available in the Google Cloud Platform Console and has Cloud SDK and the mysql client pre-installed. Cloud Shell provides a Compute Engine instance that you can use to connect to Cloud SQL.
  • Temporarily allow all IP addresses to connect to an instance. For IPv4 authorize 0.0.0.0/0 (for IPv6, authorize ::/0. This confirms that your client can connect.

Verify how you connect

If you get an error message like:

ERROR 1045 (28000): Access denied for user 'root'@'1.2.3.4' (using password: NO)

when you connect, verify that you are providing a password.

If you get an error message like:

ERROR 1045 (28000): Access denied for user 'root'@'1.2.3.4' (using password: YES)

when you connect, verify that you are using the correct password and that you are connecting over SSL if the instance requires it.

Understand connection limits

There are no QPS limits for Google Cloud SQL instances. However, there are connection, size, and App Engine specific limits in place.

Cloud SQL instances limits

Limit MySQL First Generation instances MySQL Second Generation instances PostgreSQL instances Notes
Concurrent connections Determined by tier 4,000 100
Enqueued connection requests 100 N/A N/A First Generation instances only. Incoming connection requests are briefly queued before the connection is established. The queue can accept only 100 incoming connection requests.
Storage size 250 GB Up to 10,230 GB, depending on machine type Up to 10,230 GB, depending on whether the instance has dedicated or shared vCPUs. It is possible to increase individual First Generation instance limits up to 500 GB for customers with a Silver or higher Google Cloud support package.

Google App Engine Limits

Requests from Google App Engine applications to Google Cloud SQL are subject to the following time and connection limits:

  • For apps running in the Google App Engine standard environment, all database requests must finish within the HTTP request timer, around 60 seconds. For apps running in the flexible environment, all database requests must finish within 24 hours.
  • Offline requests like cron tasks have a time limit of 10 minutes.
  • Requests to Google Cloud SQL have limitations based on the scaling type of the App Engine module and how long an instance can remain in memory (residence).
  • Each App Engine instance running in a standard environment cannot have more than 12 concurrent connections to a Google Cloud SQL instance.

Google App Engine applications are also subject to additional App Engine quotas and limits as discussed on the Quotas page.

To learn more about managing connections, see the FAQ How should I manage connections?.

Show connections and threads

If you get the “too many connections” error message or want to find out what is happening on an instance, you can show the number of connections and threads with SHOW PROCESSLIST.

From a MySQL client, run:

mysql> SHOW PROCESSLIST;
+----+-----------+--------------+-----------+---------+------+-------+----------------------+
| Id | User      | Host         | db        | Command | Time | State | Info                 |
+----+-----------+--------------+-----------+---------+------+-------+----------------------+
|  3 | user-name | client-IP    | NULL      | Query   |    0 | NULL  | SHOW processlist     |
|  5 | user-name | client-IP    | guestbook | Sleep   |    1 |       | SELECT * from titles |
| 17 | user-name | client-IP    | employees | Query   |    0 | NULL  | SHOW processlist     |
+----+-----------+--------------+-----------+---------+------+-------+----------------------+
3 rows in set (0.09 sec)

For information about how to interpret the columns returned from PROCESSLIST, see the MySQL reference.

To get a quick thread count, you can use:

mysql> SHOW STATUS WHERE Variable_name = 'Threads_connected';
+-------------------+-------+
| Variable_name     | Value |
+-------------------+-------+
| Threads_connected | 7     |
+-------------------+-------+
1 row in set (0.08 sec)

Connections from Compute Engine

If you expect that connections between your Compute Engine instance and your Cloud SQL instance will include long-lived unused connections, then you should be aware that connections with a Compute Engine instance time out after 10 minutes of inactivity. For more information, see Networking and Firewalls in the Google 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.
cat /proc/sys/net/ipv4/tcp_keepalive_time

# 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.
cat /proc/sys/net/ipv4/tcp_keepalive_time

Connecting with IPv6

If you get either of the error messages

Can't connect to MySQL server on '2001:1234::4321' (10051)
Can't connect to MySQL server on '2001:1234::4321' (101)

when you connect it is likely that you are attempting to connect to the IPv6 address of your instance but do not have IPv6 available on your workstation. You can verify whether IPv6 is functional on your workstation by going to ipv6.google.com, if it does not load then you do not have IPv6 available. To fix this you should connect to the IPv4 address or your Cloud SQL instance. You may need to add an IPv4 address to your instance first.

Instance issues

Backup

Backups for First Generation instances can fail if you have more than 10,000 database tables. For best performance, keep your number of tables to a reasonable number.

Import and export

Imports and Exports in Google Cloud SQL are the same as using the mysqldump utility except that with the Cloud SQL import/export feature, you transfer data using a Google Cloud Storage bucket. You can import and export one database, all instance databases, or selected data in CSV format.

Imports and exports into Google Cloud SQL using the import functionality (via a Cloud Storage bucket) can take a long time to complete, depending on the size of the database. This can have the following impacts:

  • On First Generation instances, operations are limited to 24 hours.
  • You cannot stop a long-running operation.

In addition, you can perform only one import or export operation at a time for each instance.

You can decrease the amount of time a single operation requires by using one of the following approaches:

  • Use the Cloud SQL import or export functionality, but do it with smaller batches of data that will take less than 24 hours to complete.

  • Don't use the Cloud SQL import or export functionality, but instead replay a dump file directly to Cloud SQL. For example, you can use cloudsql-import, which imports by replaying a mysqldump file over a MySQL connection. cloudsql-import is resilient to connection failures and instance restarts.

Other points to keep in mind when importing:

  • For First Generation instances, file system replication set to asynchronous mode is much faster than synchronous mode. For an initial import into a database, we recommend you use async mode. (You can edit an existing instance and switch between the two modes.)

  • For small instances, you can temporarily increase the tier of an instance to improve performance when importing large files.

  • Speed is greatly improved by avoiding many small transactions. This can be achieved by disabling autocommit and wrapping the entire file or large chunks of it (such as each table) in a single transaction. If your application is generating SQL files with mysqldump, you should use the --single-transaction option.

  • We recommend that you include many rows in each INSERT statement, rather than having a separate statement for every row. This can be achieved with the --extended-insert mysqldump option, which is on by default.

Disk space

If your instance reaches the maximum storage amount allowed, writes to the database fail. If you delete data, for example, by dropping a table, the space freed is not reflected in the reported Storage Used of the instance. See the FAQ How can I reclaim the space from a dropped table? for an explanation of this behavior.

Avoid data corruption

Avoid generated columns

Due to an issue in MySQL, using generated columns might result in data corruption. For more information, see MySQL bug #82736.

Clean shutdowns

When Google Cloud SQL shuts down an instance (e.g, for maintenance), no new connections are sent to the instance and existing connections are killed. The amount of time mysqld has to shutdown is capped to 1 minute. If the shutdown does not complete in that time, the mysqld process is forcefully terminated. This can result in disk writes being aborted mid-way through.

Database engines

InnoDB is the only supported storage engine for Second Generation instances because it is more resistant to table corruption than other MySQL storage engines, such as MyISAM.

By default, Cloud SQL database tables are created using the InnoDB storage engine. If your CREATE TABLE syntax includes an ENGINE option specifying a storage engine other than InnoDB, for example ENGINE = MyISAM, the table is not created and you see error messages like the following example:

ERROR 3161 (HY000): Storage engine MyISAM is disabled (Table creation is disallowed).

You can avoid this error by removing the ENGINE = MyISAM option from the CREATE TABLE command. Doing so creates the table with the InnoDB storage engine.

InnoDB is strongly recommended for First Generation instances, because of its stronger data consistency guarantees.

Changes to system tables

MySQL system tables use the MyISAM storage engine, including all tables in the mysql database, for example mysql.user and mysql.db. These tables are vulnerable to unclean shutdowns; you should issue the FLUSH CHANGES command after making any changes to these tables. If MyISAM corruption does occur, CHECK TABLE and REPAIR TABLE can get you back to good state (but not save data).

Global Transaction Identifiers (GTID)

All Second Generation instances have GTID enabled automatically. Having GTID enabled protects against data loss during replica creation and failover, and makes replication more robust. However, GTID comes with some limitations imposed by MySQL, as documented in the MySQL manual. The following transactionally-unsafe operations cannot be used with a GTID-enabled MySQL server:

  • CREATE TABLE ... SELECT statements;
  • CREATE TEMPORARY TABLE statements inside transactions;
  • Transactions or statements that affect both transactional and non-transactional tables.

If you use a transactionally-unsafe transaction, you will see an error message like the following example:

 Exception: SQLSTATE[HY000]: General error: 1786
 CREATE TABLE ... SELECT is forbidden when @@GLOBAL.ENFORCE_GTID_CONSISTENCY = 1.

Suspended state

There are a number of reasons why Google Cloud SQL may suspend an instance, including:

  • Billing issues

    For example, if the credit card for the project's billing account has expired, the instance may be suspended. You can check the billing information for a project by going to the Google Cloud Platform Console billing page, selecting the project, and viewing the billing account information used for the project.

  • Legal issues

    For example, a violation of the Google Cloud Platform Acceptable Use Policy may cause the instance to be suspended. For more information, see "Suspensions and Removals" in the Google Cloud Platform Terms of Service.

  • Operational issues

    For example, if an instance is stuck in a crash loop, i.e., it crashes while starting or just after starting, Google Cloud SQL may suspend it.

  • Replication issues

    Replication issues include cloning an instance and creating a replica. For example, suppose you create an instance to contain the recovered data from point-in-time recovery, and the recovery operation fails, then the instance (a clone) will be in a suspended state.

While an instance is suspended, you can continue to view information about it or you can delete it, if the suspension was triggered by billing issues.

Cloud SQL users with Platinum, Gold, or Silver support packages can contact our support team directly about suspended instances. All users can use the guidance above along with the google-cloud-sql forum.

Performance

Enable query logs

In order to tune the performance of your queries, you can configure Cloud SQL to log slow queries to a table. Then you can read it with a SELECT command from your client. Please note that activating the query logs will impact your instance's performance and we recommend doing it only for short period of time and not in a production environment.

You can enable logging to table by adding the MySQL flags --log_output='TABLE' and --slow_query_log=on to your instance.

Access the instance using MySQL Client and show slow queries:

SELECT * FROM mysql.slow_log;

When you are done, switch off the logging by removing the slow_query_log flag. For more information, see the MySQL reference page Selecting General Query and Slow Query Log Output Destinations.

If you are using the general_log for troubleshooting (e.g., tracking down exactly what clients are sending to mysqld), note that enabling this log can cause connection issues as described in Tips for Working with MySQL Flags.

Enable lock monitoring

InnoDB monitors provide information about the InnoDB storage engine's internal state, which you can use in performance tuning.

Access the instance using MySQL Client and obtain on-demand monitor output:

SHOW ENGINE INNODB STATUS\G

For explanations of the sections in the monitor output, see InnoDB Standard Monitor and Lock Monitor Output.

You can enable InnoDB monitors so that output is generated periodically to a file or a table, with performance degradation. For more information, see Enabling InnoDB Monitors.

Keep a reasonable number of database tables

Database tables consume system resources. A very large number can affect instance performance.

General tips

  • First Generation instances have a 10-GB limit on temporary space used by SQL operations. Some operations (for example, large aggregations with GROUP BY) could be impacted by this limit. In some cases, you can avoid hitting the temporary-space limit by writing the query differently or using different SQL syntax. Second Generation instances do not impose this limit.
  • Make sure that your machine type is sufficiently large for the workload.

For slow database inserts, updates, or deletes, consider the following actions:

  • For First Generation instances, file system replication asynchronous mode (an instance configuration setting) is much faster than synchronous mode. If you are using synchronous mode and your application can withstand some data risk then you should switch to asynchronous mode.
  • Check the locations of the writer and database; sending data a long distance introduces latency.

For slow database selects, consider the following:

  • Caching is extremely important for read performance. Compare the size of your data set to the size of RAM of your instance. Ideally, the entire data set should fit in 70% of the instance's RAM, in which case queries will not be constrained to IO performance. If this is not the case, consider increasing the size of your instance's tier.
  • If your workload consists of CPU intensive queries (sorting, regexes, other complex functions), your instance might be throttled; increase the tier.
  • Check the location of the reader and database - latency will affect read performance even more than write performance.
  • Investigate non-Cloud SQL specific performance improvements, such as adding appropriate indexing, reducing data scanned, and avoiding extra round trips.

If you observe poor performance executing queries, use EXPLAIN to identify where to:

  • Add indexes to tables to improve query performance. For example, make sure every field that you use as a JOIN key has an index on both tables.

  • Improve ORDER BY operations. If EXPLAIN shows "Using temporary; Using filesort" in the Extra column of the output, then intermediate results will be stored in a file that is then sorted, which usually results in poor performance. In this case, take one of the following steps:

    • If possible, use indexes rather than sorting. See ORDER BY Optimization for more information.

    • Increase the size of the sort_buffer_size variable for the query session.

    • Use less RAM per row by declaring columns only as large as required.

Monitor your resources on the go

Get the Google Cloud Console app to help you manage your projects.

Send feedback about...

Cloud SQL for MySQL