Seeing some weirdness with

<cfquery>

calls being run in a

<cfthread>

block (Lucee 5.3.9.166):

"ERROR","cfthread-12457","07/09/2023","21:40:10","cfthread-12457","java.lang.NullPointerException;lucee.runtime.exp.NativeException: java.lang.NullPointerException
        at lucee.runtime.listener.JavaSettingsImpl.getBundlesTranslated(JavaSettingsImpl.java:97)
        at lucee.runtime.listener.JavaSettingsImpl.getBundleDirectories(JavaSettingsImpl.java:250)
        at lucee.transformer.library.ClassDefinitionImpl.getClazz(ClassDefinitionImpl.java:110)
        at lucee.runtime.db.DataSourceSupport._initializeDriver(DataSourceSupport.java:172)
        at lucee.runtime.db.DataSourceSupport.initialize(DataSourceSupport.java:164)
        at lucee.runtime.db.DataSourceSupport.getConnection(DataSourceSupport.java:103)
        at lucee.runtime.db.DatasourceConnectionPool.loadDatasourceConnection(DatasourceConnectionPool.java:158)
        at lucee.runtime.db.DatasourceConnectionPool.getDatasourceConnection(DatasourceConnectionPool.java:119)
        at lucee.runtime.db.DatasourceManagerImpl.getConnection(DatasourceManagerImpl.java:82)
        at lucee.runtime.tag.Query.executeDatasoure(Query.java:1119)
        at lucee.runtime.tag.Query._doEndTag(Query.java:699)
        at lucee.runtime.tag.Query.doEndTag(Query.java:565)
        at wheels.model.adapters.base_cfc$cf.udfCall2(/wheels/model/adapters/Base.cfc:302)

(

Base.cfc:302

is CFWheels's catch-all

<cfquery>

tag for its "orm"). This is cropping up in our logs every few hours, from a process that runs maybe a dozen queries within as many separate threads. The process runs every 15min. We pushed this logic back into separate threads a week or so ago. Prior to that the equiv single-threaded code had been running quite happily every 15min for a few years without issue. Looking at that section of the stack trace, it looks to me like Lucee is reloading its DB drivers for some reason, and poss more than one thread is doing so at a time, and Lucee is not blocking the reload code properly? Just a wild guess. Had look through Jira and did not find anything similar. Thoughts?

my vague memory is this has been fixed since

Ah ok, that's good news. For everyone else, anyhow.