Database Connectors: faulty CustomDriver registration logic
Mantenedores costumam responder em até 2 dias
Ninguém assumiu esta issue ainda.
Avaliação
- Dificuldade
- 5/5
- Tempo estimado
- Mais de uma semana
- Facilidade para iniciantes
- 28/100
Direção de pesquisa
Start at CustomDriver.java around the DriverManager deregister path (the linked line ~130) and how delegate is used in connect. Reproduce with an Oracle JDBC driver loaded as a Resource rather than server-lib/custom-lib, after the PR #340 behavior. Decide whether deregister must be skipped or isolated for drivers whose deregister hook runs global cleanup (OracleDriver’s TimeoutInterruptHandler.stopTimer / Clock static init). Done means Oracle Database Reader channels can start more than once without NoClassDefFoundError on oracle.net.nt.Clock, without forcing the jar into server-lib.
Escrita pelo modelo de indexação a partir do texto da issue.
Descrição
From testing the fix submitted with PR #340 (more context in this Discord thread), it seems like there might be a problem with logic in CustomDriver initialisation, specifically when used with Oracle JDBC driver.
See this line: CustomDriver.java:130
Intendedly, in order to prevent default first-match driver selection behavior of DriverManager, CustomDriver attempts to deregister the configured driver from DriverManager and use it directly through the delegate property.
However, in the case of OracleDriver (and perhaps others), the deregister function it provides when registering with DriverManager contains globally destructive cleanup code which prevents the class from being used again after deregistering.
Specifically, from the ojdbc17-23.26.2.0.0.jar, when this deregister function of OracleDriver.java gets called as part of CustomDriver init:
private static void deregister() {
unRegisterMBeans();
AbstractTrueCacheConnectionPools.destroyAllPools();
OracleTimeoutThreadPerVM.stopWatchdog();
ThreadedCachingBlockSource.stopBlockReleaserThread();
TimeoutInterruptHandler.stopTimer();
if (threadPool != null) {
threadPool.shutdownNow();
}
HAManager.shutdownAll();
OracleDataSource.cleanup();
TxnFailoverManagerImpl.cleanup();
TcpMultiplexer.stop();
}
The TimeoutInterruptHandler.stopTimer(); call causes the static init block of oracle.net.nt.Clock to fail on future attempts to use the OracleDriver class:
package oracle.net.nt;
public class Clock {
...
static {
if (cachedCurrentTimeMillisTaskStarted.compareAndSet(false, true)) {
TimeoutInterruptHandler.scheduleAtFixedRate(cacheCurrentTimeMillis, 0L, ACCURACY_IN_MILLIS);
}
}
}
Resulting in the following error when attempting to Start a deployed channel with Database Reader source of Oracle type (imported as Resource not part of custom-lib directory):
[2026-07-15 21:02:42.385] ERROR (com.mirth.connect.server.channel.ErrorTaskHandler:26): com.mirth.connect.donkey.server.StartException: Failed to start channel test db connection 2 (5e6adfe6-4200-4543-bfb1-01dcd7c33969).
at com.mirth.connect.donkey.server.channel.Channel.start(Channel.java:792)
at com.mirth.connect.server.controllers.DonkeyEngineController$DeployTask.doDeploy(DonkeyEngineController.java:2031)
at com.mirth.connect.server.controllers.DonkeyEngineController$DeployTask.execute(DonkeyEngineController.java:1835)
at com.mirth.connect.server.channel.ChannelTask.call(ChannelTask.java:67)
at com.mirth.connect.server.channel.ChannelTask.call(ChannelTask.java:16)
at java.base/java.util.concurrent.FutureTask.run(Unknown Source)
at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source)
at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
at java.base/java.lang.Thread.run(Unknown Source)
Caused by: java.lang.NoClassDefFoundError: Could not initialize class oracle.net.nt.Clock
at oracle.net.nt.NetStatImpl.incrementBytesSent(NetStatImpl.java:80)
at oracle.net.nt.TimeoutSocketChannel.logWrite(TimeoutSocketChannel.java:887)
at oracle.net.nt.TimeoutSocketChannel.doBlockedWrite(TimeoutSocketChannel.java:833)
at oracle.net.nt.TimeoutSocketChannel.write(TimeoutSocketChannel.java:772)
at oracle.net.nt.SSLSocketChannel.writeToSocket(SSLSocketChannel.java:931)
at oracle.net.nt.SSLSocketChannel.shutdown(SSLSocketChannel.java:443)
at oracle.net.nt.SSLSocketChannel.disconnect(SSLSocketChannel.java:991)
at oracle.net.nt.TcpNTAdapter.disconnect(TcpNTAdapter.java:636)
at oracle.net.ns.NSProtocolNIO.establishConnectionAfterRefusePacket(NSProtocolNIO.java:584)
at oracle.net.ns.NSProtocolNIO.handleConnectPacketResponse(NSProtocolNIO.java:482)
at oracle.net.ns.NSProtocolNIO.negotiateConnection(NSProtocolNIO.java:287)
at oracle.net.ns.NSProtocol.connect(NSProtocol.java:369)
at oracle.jdbc.driver.T4CConnection.connectNetworkSessionProtocol(T4CConnection.java:3978)
at oracle.jdbc.driver.T4CConnection.logon(T4CConnection.java:1187)
at oracle.jdbc.driver.PhysicalConnection.connect(PhysicalConnection.java:1236)
at oracle.jdbc.driver.T4CDriverExtension.getConnection(T4CDriverExtension.java:107)
at oracle.jdbc.driver.OracleDriver.connect(OracleDriver.java:817)
at oracle.jdbc.driver.OracleDriver.connect(OracleDriver.java:720)
at com.mirth.connect.connectors.jdbc.CustomDriver.connect(CustomDriver.java:153)
at com.mirth.connect.connectors.jdbc.CustomDriver.connect(CustomDriver.java:148)
at com.mirth.connect.connectors.jdbc.DatabaseReceiverQuery.initSelectConnection(DatabaseReceiverQuery.java:338)
at com.mirth.connect.connectors.jdbc.DatabaseReceiverQuery.initConnection(DatabaseReceiverQuery.java:101)
at com.mirth.connect.connectors.jdbc.DatabaseReceiverQuery.start(DatabaseReceiverQuery.java:87)
at com.mirth.connect.connectors.jdbc.DatabaseReceiver.onStart(DatabaseReceiver.java:115)
at com.mirth.connect.donkey.server.channel.SourceConnector.start(SourceConnector.java:106)
at com.mirth.connect.donkey.server.channel.PollConnector.start(PollConnector.java:31)
at com.mirth.connect.donkey.server.channel.Channel.start(Channel.java:768)
... 8 more
Caused by: java.lang.ExceptionInInitializerError: Exception java.lang.IllegalStateException: Timer already cancelled. [in thread "Channel DeployTask Thread on (5e6adfe6-4200-4543-bfb1-01dcd7c33969) < pool-8-thread-1"]
at java.base/java.util.Timer.sched(Unknown Source)
at java.base/java.util.Timer.scheduleAtFixedRate(Unknown Source)
at oracle.net.nt.TimeoutInterruptHandler.scheduleAtFixedRate(TimeoutInterruptHandler.java:303)
at oracle.net.nt.Clock.<clinit>(Clock.java:58)
at oracle.net.nt.NetStatImpl.incrementBytesSent(NetStatImpl.java:80)
at oracle.net.nt.TimeoutSocketChannel.logWrite(TimeoutSocketChannel.java:887)
at oracle.net.nt.TimeoutSocketChannel.doBlockedWrite(TimeoutSocketChannel.java:833)
at oracle.net.nt.TimeoutSocketChannel.write(TimeoutSocketChannel.java:772)
at oracle.net.nt.SSLSocketChannel.writeToSocket(SSLSocketChannel.java:931)
at oracle.net.nt.SSLSocketChannel.wrapHandshakeMessage(SSLSocketChannel.java:699)
at oracle.net.nt.SSLSocketChannel.doSSLHandshake(SSLSocketChannel.java:539)
at oracle.net.nt.SSLSocketChannel.write(SSLSocketChannel.java:199)
at oracle.net.ns.NIOPacket.writeToSocketChannel(NIOPacket.java:378)
at oracle.net.ns.NIOConnectPacket.writeToSocketChannel(NIOConnectPacket.java:291)
at oracle.net.ns.NSProtocolNIO.negotiateConnection(NSProtocolNIO.java:214)
... 24 more
Note that the channel could only be deployed once the fix from PR #340 is in place. Without it, channel deploy fails as documented in #338.
The simple workaround for this is to keep the driver libraries within server-lib. However, this defeats the stated purpose of CustomDriver to allow multiple different versions of the same Driver class to be used.
Is this a fundamental flaw in CustomDriver's expectation to be able to trigger driver deregistering, or could there be some straightforward way to accomodate such drivers with terminal cleanup functions?
- Linguagem predominante
- Java
- Estrelas
- 227
- Forks
- 81
- Merge médio
- 3d 6h
- PRs com merge (30d)
- 15
Preparar o ambiente
- Inclui um Dockerfile ou arquivo Docker Compose
- Sem modelo de pull request
- Ler o guia de contribuição
Primeiros passos
- Leia a issue inteira e depois o guia de contribuição do projeto.
- Comente na issue dizendo que vai assumir — evita que duas pessoas façam o mesmo trabalho.
- Faça um fork do repositório e trabalhe em uma branch.
- Abra um pull request que referencie o número da issue.
Mais de OpenIntegrationEngine/engine
-
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 68/100
OpenIntegrationEngine/engine#485 ·
Mantenedores costumam responder em até 2 dias
-
enhancement
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 62/100
OpenIntegrationEngine/engine#478 · 1 comentário ·
Mantenedores costumam responder em até 2 dias
-
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 68/100
OpenIntegrationEngine/engine#355 ·
Mantenedores costumam responder em até 2 dias
-
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 68/100
OpenIntegrationEngine/engine#347 ·
Mantenedores costumam responder em até 2 dias
-
Fix call-home in test runAberta
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 68/100
OpenIntegrationEngine/engine#313 ·
Mantenedores costumam responder em até 2 dias
Todas as issues de OpenIntegrationEngine/engine
Issues semelhantes
-
pdfium-drift
Dificuldade 1/5 Menos de uma hora Facilidade para iniciantes 82/100
Stirling-Tools/JPDFium#176 ·
Mantenedores costumam responder em até 1 dia
-
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 72/100
hiero-ledger/hiero-consensus-node#27712 ·
Mantenedores costumam responder em até 1 dia
-
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 61/100
Vakore/ZappierGames#162 ·
-
Improve README documentation clarity and fix minor wording issuesTalvez já em andamento @simplesilence assumiu hoje. Aberta
Dificuldade 1/5 1-3 horas Facilidade para iniciantes 84/100
resilience4j/resilience4j#2567 ·
Mantenedores costumam responder em até 6 dias
-
Priority/High
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 65/100
wso2/dpdp-accelerator#417 ·
Mantenedores costumam responder em até 1 dia