Running ORMReload often causes timeout or Lucee to crash

Box Modules cborm
1

I am running Lucee 6.0.0.585

I am using a mix of Coldbox ORM and the qb module. In my Application.cfc I have this bit of logic that allows me to do an ORMReload. Sometimes it works fine when I trigger this by adding a query param to the URL as in ?fwreinit=1&refresh=1 for example, but it times out and then I have to restart Lucee.

if (	!isNull( url.rebuild ) ||	!isNull( url.refresh )	) {
		ORMReload();
    }

error is: public_html/index.cfm has run into a timeout (timeout: 50 seconds) and has been stopped. The thread started 58500ms ago. Open locks at this time [640bc157f893e055bb614b9eeee79693, module16886645ed666b41e092d593f18e04f5.activation.cbsecurity].

coldbox/system/ioc/Builder.cfc: line 591

589: targetObject = “”
590: ){
591: var refLocal = {};
592: var DSLNamespace = listFirst( arguments.definition.dsl, “:” );
593:
called from /home/public_html/coldbox/system/ioc/Injector.cfc: line 1289
called from /home/public_html/coldbox/system/ioc/Injector.cfc: line 915
called from /home/public_html/coldbox/system/ioc/scopes/Singleton.cfc: line 92
called from /home/public_html/coldbox/system/ioc/Injector.cfc: line 581
called from /home/public_html/coldbox/system/web/services/InterceptorService.cfc: line 484

Does this ever happen to anyone else, should I reload another way? I don’t have tons of model objects either.

2

Does this ever happen to anyone else, should I reload another way?

Yes and no. Yes, I’ve seen tremendous performance costs incurred when running an ORMReload, and no, I don’t know of any better reload method.

Are you using the Ortus ORM extension, or the old Lucee Hibernate extension? We’ve dramatically improved ORM reload times in the Ortus ORM extension - see the docs on v5.4.29.22, where we improved ORMReload time by up to 60%, and the v6.0.0 docs, where we document further ORMReload improvements.

I can’t debug that exception for you, since I don’t have 1) your app, or 2) the full stack trace in front of me, but in general I would follow this process to debug:

  1. Start up Fusion Reactor and profile the ormReload request. Dig into any slow code. This helped me discover LDEV-4064 several years ago.
  2. Install the Ortus ORM extension and benefit immediately frometter docs](https://orm-extension.ortusbooks.com/) and faster startup times.
  3. Rename all but a single entity to *.txt. Run the ORMReload andbug. Then rename back to .cfc until you start seeing the ORMReload slowdown
  4. Look at all code in the pseudo-constructor of each entity. You want to avoid doing any complex work in the pseudo-constructor, since I believe Lucee has to build at least that much of the component just to grab entity metadata.

With that said… your issue seems more like a CBSecurity configuration issue than an ORMReload one. Of course, I can’t say that for sure without seeing, I dunno, the full exception and stack trace, the CBSecurity version, coldbox version, and your coldbox and cbsecurity settings. :slight_smile:

3

Thanks for the detailed reply @MichaelBorn.

I am using the “cborm” extension shown here: https://coldbox-orm.ortusbooks.com/

It sounds like the extension you’re referring to is different than what I’m using, but please confirm.

Your suggestions around debugging make sense. I will look at my entities and see if I can isolate what is causing the issue. In some ways, I hope I can consistently reproduce it.

I also wonder if it’s worth increasing the request timeout, at least in dev to see if that helps while still digging into solving the underlying issue.

Thanks for your support!

4

Yes, that’s a ColdBox extension. You also need a Lucee extension which bundles Hibernate. You have two choices: the Lucee Hibernate extension (which is no longer supported or maintained) or the Ortus ORM Extension.