I am using the standard HikariCP implementation in my SpringBoot 2.0.1 application. But after a while I get the same error over and over
HikariPool-1 – Connection is not available, request timed out after 30000ms
I first checked the code if there are any not closed connection or missing Transactional annotations but I did not find anything. I also tried then to increase the pool and decrease the time out in my application.yml but it seems that this does not have any effect.
The weird thing is that HikariCP seems to create only 4 pooled connection and I am not able to override these properties.
HikariPool-1 - Timeout failure stats (total=4, active=4, idle=0, waiting=100)
This is my application.yml file
spring:
jpa:
hibernate:
ddl-auto: update
datasource:
hikari:
maximum-pool-size: 64
connection-timeout: 5000
Your application is not able to get the connection within the 30s which is the default connection timeout for HikariCP. There can be reasons for that such as:
The Pool size is small, and the requests take too long to execute (more than 30 s).
Possible explanation is that for e.g your pool size is 1 and your one request/query takes approximately 500ms to execute. suppose you have opened for example approximately 200 concurrent query requests, then there might be a chance that some requests will be timed out. (one request has waited for the 30s but a connection was not available for it).
Disclaimer: Numbers are taken randomly to explain and is not tested.
To solve this you might want to increase the connectionTimeout or increase the pool size. ( Also check for the memory requirements).
In application.properties set the values
spring.datasource.hikari.minimumIdle=10
spring.datasource.hikari.maximumPoolSize=10
spring.datasource.hikari.idleTimeout=120000
spring.datasource.hikari.maxLifetime=500000
Official Hikari CP documentation link
There can be another reason for this is that your connection is not getting closed properly use try-with-resources to close your connection
e.g
try(Connection c = dataSource.getConnection()) {
//ToDo: Add your logic statement here
log.debug("[d] Tenant '{}' added.", tenantId);
}
Related
I am trying to check the JDBC connection status after each thread execution whether it is close or open?.
In my thread group there are three things
JDBC connection configuration
JDBC request (select * from employee)
JSR223 PostProcessor
Script :
def connection = org.apache.jmeter.protocol.jdbc.config.DataSourceElement.getConnection('ConnectionString')
log.info('*************Connection closed: '+ connection.isClosed())
Above script is logging the connection status after each thread execution when loop count is 1. Problem here as soon as I change the loop count to >= 2. it started throwing the error below error
Problem in JSR223 script, JSR223 PostProcessor
javax.script.ScriptException: java.sql.SQLException: Cannot get a connection, pool error Timeout waiting for idle object
And when I remove the Post processor and increase the loop count it is working fine.
Logs :
2023-02-19 16:15:33,599 INFO o.a.j.t.JMeterThread: Thread started: DB Thread Group 2-1
2023-02-19 16:15:38,054 DEBUG o.a.j.p.j.AbstractJDBCTestElement: executing jdbc:SELECT * FROM EMPLOYEE
2023-02-19 16:15:38,623 INFO o.a.j.e.J.JSR223 PostProcessor: *************Connection closed: false
2023-02-19 16:15:58,637 ERROR o.a.j.e.JSR223PostProcessor: Problem in JSR223 script, JSR223 PostProcessor
javax.script.ScriptException: java.sql.SQLException: Cannot get a connection, pool error Timeout waiting for idle object, borrowMaxWaitDuration=PT10S
at org.codehaus.groovy.jsr223.GroovyScriptEngineImpl.eval(GroovyScriptEngineImpl.java:320) ~[groovy-jsr223-3.0.11.jar:3.0.11]
at org.codehaus.groovy.jsr223.GroovyCompiledScript.eval(GroovyCompiledScript.java:71) ~
In the JDBC configuration are you using a connection pool to request connections?
What your test shows is that the JSR223 script is closing the connection, which is probably a good thing from a coding perspective, but the next iteration of the loop tries to execute a request with a closed Connection and blammo. If you switch from raw connections to a connection pool when the JSR 223 closes the connection it'll be returned to the pool and remain open for the next iteration of the loop. You'll have to switch to using DataSource API typically for this, but it's a minor tweak to the script.
I can think of 2 possible reasons:
Either your database is down/overloaded/not reachable via JDBC
Or your connection pool settings need to be tweaked, i.e. max number of connections and/or wait time need to be increased:
In general I don't think your approach is correct, as per JavaDoc:
This method generally cannot be called to determine whether a connection to a database is valid or invalid. A typical client can determine that a connection is invalid by catching any exceptions that might be thrown when an operation is attempted.
So you might want to increase debug logging verbosity for JMeter, your JDBC driver and Java SQL namespace instead
I was facing this issue for my springboot application that connects to a DB and MQ, and uses Atomikos Transaction manager.
com.atomikos.jms.AtomikosJMSException|Connection pool exhausted - try increasing 'maxPoolSize' and/or 'borrowConnectionTimeout' on the AtomikosConnectionFactoryBean.
com.atomikos.datasource.pool.PoolExhaustedException: ConnectionPool: pool is empty - increase either maxPoolSize or borrowConnectionTimeout
at com.atomikos.datasource.pool.ConnectionPool.waitForAtLeastOneAvailableConnection(ConnectionPool.java:326)
at com.atomikos.datasource.pool.ConnectionPool.findOrWaitForAnAvailableConnection(ConnectionPool.java:144)
at com.atomikos.datasource.pool.ConnectionPool.borrowConnection(ConnectionPool.java:132)
at com.atomikos.datasource.pool.ConnectionPoolWithSynchronizedValidation.borrowConnection(ConnectionPoolWithSynchronizedValidation.java:23)
at com.atomikos.jms.AtomikosConnectionFactoryBean.createConnection(AtomikosConnectionFactoryBean.java:601)
at org.springframework.jms.support.JmsAccessor.createConnection(JmsAccessor.java:196)
at org.springframework.jms.listener.AbstractPollingMessageListenerContainer.access$100(AbstractPollingMessageListenerContainer.java:77)
at org.springframework.jms.listener.AbstractPollingMessageListenerContainer$MessageListenerContainerResourceFactory.createConnection(AbstractPollingMessageListenerContainer.java:490)
at org.springframework.jms.connection.ConnectionFactoryUtils.doGetTransactionalSession(ConnectionFactoryUtils.java:325)
at org.springframework.jms.listener.AbstractPollingMessageListenerContainer.doReceiveAndExecute(AbstractPollingMessageListenerContainer.java:281)
at org.springframework.jms.listener.AbstractPollingMessageListenerContainer.receiveAndExecute(AbstractPollingMessageListenerContainer.java:245)
at org.springframework.jms.listener.DefaultMessageListenerContainer$AsyncMessageListenerInvoker.invokeListener(DefaultMessageListenerContainer.java:1189)
at org.springframework.jms.listener.DefaultMessageListenerContainer$AsyncMessageListenerInvoker.executeOngoingLoop(DefaultMessageListenerContainer.java:1179)
at org.springframework.jms.listener.DefaultMessageListenerContainer$AsyncMessageListenerInvoker.run(DefaultMessageListenerContainer.java:1076)
at java.lang.Thread.run(Thread.java:748)
I tried printing the maxPoolSize and found that it is 1. This page came across in between (https://www.atomikos.com/Documentation/ConfiguringJms) and I found the line where they increased the MaxPoolSize to 5. I just tried setting it to 2 and it worked.
AtomikosConnectionFactoryBean xaConnectionFactory = new AtomikosConnectionFactoryBean();
xaConnectionFactory.setXaConnectionFactory(ibmMQXAConnectionFactory);
xaConnectionFactory.setMaxPoolSize(2);
Can someone help me to understand what should be the ideal poolsize. what it is for etc?
In order to process messages Atomikos uses DB and JMS connections (in your case).
These connections are taken from the pools of available connections. To get the idea why connection pools are needed, please follow this link as a starting point - Connection_pool
To put it simple - in order to process one message at a time Atomikos needs one DB and one JMS connection/session. So if you plan to process 10 messages in parallel, each connection pool size must be at least 10 (10 for DB and 10 for JMS connection pools respectively).
Does HikariCP supports command timeout in Spring Boot application similar to C#
I am using Hikari Connection Pool in my Spring boot application. I have enabled connectionTimeout using the following configuration
spring.datasource.hikari.connectionTimeout: 30000
If I increase the number of concurrent users I'm getting the following exception in logs
Caused by: org.springframework.jdbc.CannotGetJdbcConnectionException: Failed to obtain JDBC Connection;
nested exception is java.sql.SQLTransientConnectionException: HikariPool-1 - Connection is not available, request timed out after 30000ms.
I am perfectly find with the above exception. I can increase the number of connections. But my concern is, there are few endpoints which are taking more than 2 minutest to respond. Those endpoints got DB Connection from pool but took more time to process. Is there a setting in which I can mentioned some timeout so that if DB takes more than some time (Operation time -Say 40 seconds) then it should send SQL exception. Similar to command timeout in C#
Application.properties
# A list of all Hikari parameters with a good explanation is available on https://github.com/brettwooldridge/HikariCP#configuration-knobs-baby
# This property controls the minimum number of idle connections that HikariCP tries to maintain in the pool. Default: same as maximumPoolSize
spring.datasource.hikari.minimumIdle: 10
# This property controls the maximum size that the pool is allowed to reach, including both idle and in-use connections. Basically this value will determine the maximum number of actual connections to the database backend.
# Default: 10
spring.datasource.hikari.maximumPoolSize: 20
#This property controls the maximum number of milliseconds that a client (that's you) will wait for a connection from the pool. If this time is exceeded without a connection becoming available, a SQLException will be thrown.
#Lowest acceptable connection timeout is 250 ms. Default: 30000 (30 seconds)
spring.datasource.hikari.connectionTimeout: 30000
# This property controls the maximum amount of time that a connection is allowed to sit idle in the pool. This setting only applies when minimumIdle is defined to be less than maximumPoolSize
# Default: 600000 (10 minutes)
spring.datasource.hikari.idleTimeout: 600000
# This property controls the maximum lifetime of a connection in the pool. An in-use connection will never be retired, only when it is closed will it then be removed.
# Default: 1800000 (30 minutes)
spring.datasource.hikari.maxLifetime: 1800000
# This property sets a SQL statement that will be executed after every new connection creation before adding it to the pool. Default: none
spring.datasource.hikari.connectionInitSql: SELECT 1 FROM DUAL
As explained in this comment, seems that connectionTimeout value is used for acquire connection timeout.
You can try setting that to a higher value than 30 secs and see if that helps.
Druid connection pool have this setting that allows you to set up the timeout:
spring.datasource.druid.query-timeout=10000
I Have the following Hikari configuration in my spring boot app. Queries are taking more than the connection-timeout time set. However, timeout never happened. I am keeping as low as possible to simulate the connection timeout.
HikariConfig dataSourceConfig = new HikariConfig();
dataSourceConfig.setDriverClassName(config.driver);
dataSourceConfig.setJdbcUrl(config.url);
dataSourceConfig.setUsername(config.user);
dataSourceConfig.setPassword(config.password);
dataSourceConfig.setConnectionTestQuery(config.validationQuery);
dataSourceConfig.setMaximumPoolSize(config.poolMax);
dataSourceConfig.setConnectionTimeout(300);
dataSourceConfig.setIdleTimeout(10000);
dataSourceConfig.setMaxLifetime(60000);
JdbcTemplate jdbcTemplate = new JdbcTemplate(new HikariDataSource(dataSourceConfig));
Here is some log that shows the query ran for more than 300ms.
Time elapsed to run query......2913
Using Hikari 3.2 and mariadb
Thanks.
From: https://github.com/brettwooldridge/HikariCP
connectionTimeout
This property controls the maximum number of milliseconds that a client (that's you) will wait for a connection from the pool. If this time is exceeded without a connection becoming available, a SQLException will be thrown. Lowest acceptable connection timeout is 250 ms. Default: 30000 (30 seconds)
So this property is more about how long your application will wait for a connection, not how long a query is allowed to execute.
I think what you want is "max_statement_time": https://mariadb.com/kb/en/library/server-system-variables/#max_statement_time
Maximum time in seconds that a query can execute before being aborted. This includes all queries, not just SELECT statements, but excludes statements in stored procedures. If set to 0, no limit is applied.
HikariPool-1 - Connection marked as broken because of SQLSTATE(08S01), ErrorCode(-99999)
java.sql.SQLNonTransientConnectionException: Communication link failure. (Read timed out)
I kept getting this problem, when creating a table with data, which was about 6million records. On ACS, an sql developer, it took around 1½ minute.
I found out that default connection validation (spring.datasource.hikari.validationTimeout) was set to 5000 (5 seconds).
This lead to hikari, not being able to validate the connection, because it was busy for long time.
Solution was setting this property to a high enough number (i set it to 5minutes - 300000)