This page contains a list of the most frequent issues you might run intowhen working with Cloud SQL instances and steps you can take to addressthem. Also review theKnown issues,Troubleshooting, andSupport page pages.

View logs

To see information about recent operations, you can view theCloud SQL instance operation logsor theMySQL error logs.

Instance unresponsive

If your instance stops responding to connections or performance is degraded,make sure it conforms to theOperational Guidelines. If it does notconform to these guidelines, it is not covered bytheCloud SQL SLA.

Connection issues

See theDebugging connectionissues page or theConnectivity section in the troubleshooting page for help with connectionproblems.

Instance issues

Backups

For the best performance forbackups, keep thenumber of tables to a reasonable number.

For other backups issues, see theBackups section in the troubleshooting page.

Import and export

Imports and exports in Cloud SQL are thesame as using themysqldump utility, except that with the Cloud SQLimport/export feature, you transfer data using a Cloud Storage bucket.

Imports into Cloud SQL and exports out of Cloud SQL can take a long time to complete, depending on the size of the data being processed. This can have the following impacts:

• You can't stop a long-running Cloud SQL instance operation.
• You can perform only one import or export operation at a time for each instance, and a long-running import or export blocks other operations, such as daily automated backups. Serverless exports allow you to run other operations, including editing instances, import, failover, and unblocking daily automated backups.

You can decrease the amount of time it takes to complete each operation by using the Cloud SQL import or export functionality with smaller batches of data.

For exports, you can perform the export from aread replica or useserverless export to minimize the impact on database performance and allow other operations to run on your instance while an export is running.

Other points to keep in mind when importing:

• If your import is crashing, it could be due to an out-of-memory (OOM) error.If this is the case, you can try using MySQL commands directly to add the--extended-insert=FALSE --complete-insert parameters. These parameters reducethe speed of your import, but also reduce the amount of memory the importrequires.

For other import and export issues, see theImport and export section in the troubleshooting page.

Disk space

Reaching the maximum storage limit can also cause the instance to get stuck inrestart.

Avoid data corruption

Avoid generated columns

Due to an issue in MySQL, using generated columns might result in datacorruption. For more information, seeMySQL bug #82736.

Clean shutdowns

When Cloud SQL shuts down an instance (e.g, for maintenance), no new connectionsare sent to the instance and existing connections are ended. The amount oftime mysqld has to shutdown is capped to 1 minute. If the shutdown does notcomplete in that time, the mysqld process is forcefully stopped. This canresult in disk writes being aborted mid-way through.

Database engines

InnoDB is the only supported storage engine for MySQL instancesbecause it is more resistant to table corruption than other MySQL storageengines, such asMyISAM.

By default, Cloud SQL database tables are created using the InnoDB storageengine. If yourCREATE TABLE syntax includes anENGINEoption specifying a storage engine other than InnoDB, for exampleENGINE = MyISAM, the table is not created and you see error messageslike the following example:

ERROR3161(HY000):StorageengineMyISAMisdisabled(Tablecreationisdisallowed).

You can avoid this error by removing theENGINE = MyISAM option from theCREATE TABLE command. Doing so creates the table with the InnoDB storageengine.

Changes to system tables

MySQL system tables use the MyISAM storage engine, including all tables in themysql database, for examplemysql.user andmysql.db. These tables arevulnerable to unclean shutdowns; issue theFLUSH CHANGES commandafter making changes to these tables. If MyISAM corruptiondoes occur,CHECK TABLE andREPAIR TABLE can get you back to good state(but not save data).

Global Transaction Identifiers (GTID)

All MySQL instances have GTID enabled automatically. Having GTIDenabled protects against data loss during replica creation and failover, andmakes replication more robust. However, GTID comes with some limitationsimposed by MySQL, as documented in theMySQL manual.The following transactionally unsafe operations cannot be used with aGTID-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 see an error messagelike the following example:

Exception:SQLSTATE[HY000]:Generalerror:1786CREATETABLE...SELECTisforbiddenwhen@@GLOBAL.ENFORCE_GTID_CONSISTENCY=1.

Work with triggers and stored functions

If your instance has binary logging enabled, and you need to work withtriggers or stored functions, make sure your instance has thelog_bin_trust_function_creators flagset toon.

Suspended state

There are various reasons why Cloud SQL may suspend an instance,including:

• Billing issues

  For example, if the credit card for the project's billing account hasexpired, the instance may be suspended. You can check the billinginformation for a project by going to the Google Cloud consolebilling page, selecting the project, and viewingthe billing account information used for the project. After you resolvethe billing issue, the instance returns to runnable status withina few hours.

• Key issues with Cloud Key Management Service

  For example, if the key version of the Cloud KMS that's used to encrypt the user data in the Cloud SQL instance isn't present, access to the key is revoked, or if the key is deactivated, deleted or unreachable for several hours after multiple retries. For more information, seeUsing customer-managed encryption keys (CMEK).

• Legal issues

  For example, a violation of theGoogle Cloud Acceptable Use Policy may cause theinstance to be suspended. For more information, see "Suspensions andRemovals" in theGoogle Cloud Terms of Service.

• Operational issues

  For example, if an instance is stuck in a crash loop (it crasheswhile starting or just after starting), Cloud SQL may suspend it.

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

Cloud SQL users with Platinum, Gold, or Silversupport packages can contact our support team directly aboutsuspended instances. All users can use the earlier guidance along with thegoogle-cloud-sql forum.

Performance

Overview

Cloud SQL supports performance-intensive workloads with up to 60,000 IOPSand no extra cost for I/O. IOPS and throughput performance depends on disk size,instance vCPU count, and I/O block size, among other factors.

Your instance's performance also depends on yourchoice of storage type andworkload.

Learn more about:

• Persistent disks and performance.
• Performance and throttling metrics.
• Optimizing disk performance.
• Other factors that affect performance.

Enable query logs

To tune the performance of your queries, you can configureCloud SQL to log slow queries byadding the database flags--log_output='FILE' and--slow_query_log=on to your instance.This makes the log output available using theLogs Viewer in the Google Cloud console.Note thatGoogle Cloud Observability logging charges apply.

Do not set log_output toTABLE. Doing so can cause connection issues asdescribed inTips for working with flags.

You can refer tothis tutorialfor instructions to log and monitor Cloud SQL for MySQL slow queries usingCloud Logging and Monitoring.

Enable lock monitoring

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

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

SHOWENGINEINNODBSTATUS\G

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

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

Use performance schema

TheMySQL Performance Schema is afeature for monitoring MySQL Server execution at a low level. The most accessibleway to consume the stats generated in performance_schema is throughMySQL WorkbenchPerformance Reports functionality.

Keep a reasonable number of database tables

Database tables consume system resources. A large numbercan affect instance performance and availability, and cause the instance tolose its SLA coverage.Learn more.

General performance tips

• Check the locations of the writer and database; sending data a long distanceintroduces latency.

For slow database selects, consider the following:

• Caching is important for read performance. Compare the size of yourdataset to the size of RAM of your instance. Ideally, the entire datasetfits within 70% of the instance's RAM, in which case queries are notconstrained to IO performance. If not, consider increasing thesize of your instance's RAM.
• If your workload consists of CPU intensive queries (sorting, regularexpressions, other complex functions), your instance might be throttled;increase the vCPUs.

If you observe poor performance executing queries, useEXPLAIN. EXPLAIN is a statement you add toother statements, like SELECT, and it returns information about how MySQLexecutes the statement. It works with SELECT, DELETE, INSERT, REPLACE, andUPDATE. For example,EXPLAIN SELECT * FROM myTable;.

UseEXPLAIN to identify where you can:

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

• ImproveORDER BY operations. IfEXPLAIN shows"Using temporary; Using filesort" in theExtra column of the output, thenintermediate results are stored in a file that is then sorted, whichusually results in poor performance. In this case, take one of the followingsteps:

  • If possible, use indexes rather than sorting. SeeORDER BY Optimization formore information.

  • Increase the size of thesort_buffer_sizevariable for the query session.

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

Troubleshoot

For other Cloud SQL issues, see thetroubleshooting page.

Error messages

For specific API error messages, see theError messages reference page.

Troubleshoot customer-managed encryption keys (CMEK)

Cloud SQL administrator operations, such as create, clone, or update, might fail due to Cloud KMS errors, and missing roles or permissions. Common reasons for failure include a missing Cloud KMS key version, a disabled or destroyed Cloud KMS key version, insufficient IAM permissions to access the Cloud KMS key version, or the Cloud KMS key version is in a different region than the Cloud SQL instance. Use the following troubleshooting table to diagnose and resolve common problems.

Customer-managed encryption keys troubleshooting table