I am not sure what it's needing to phone home for . We install the extensions (copy lex files to the lucee-server/context/deploy folder) and then do a warmup in the docker build phase. This warms up everything and downloads all the requirements as that is what it needs to do to install the extensions AFAIK. I havent found it "erratic and unreliable" in the 1000's of deployments we have done so far using this.

what it's needing to phone home for

It's phoning home when the stuff ISN'T warmed. ie: to get a proper functional and complete installation of Lucee and its CFML installed, there are two options (as far as I can tell): • prewarm it somehow (under discussion here) • just let it phone home when it tries to run some code calling

encodeForHtml

or

imageNew

or [fuck knows, I don't know what Lucee has decided is an "extentions" for these purposes] The erraticness I mentioned was off the back of Jamie seeming to indicate that the pre-warm thing didn't seem very reliable. I think it was you who suggested the prewarm process needed to have small apps that actually hit the functionality in question so Lucee went and got the extensions? To do that, I'd need an app that went through and hit every bit of CFML functionality that's in extensions? Am I understanding this correctly? So it's not just a "call something like

/opt/lucee/bin/prewarm

and it will go finish its install", it's just a way of running Lucee quickly to run some code that then triggers the phone-home stuff?

I see the issue here. I think this is the encodeforHTML specifically asking for that, rather than any other functionality correct?

You need to set the server deploy.log to trace to see what happens under the hood. Not sure if you saw my correction but I no longer recommend that prewarm method. It doesn't work as of last September. Use this instead, simply:

prewarm.sh

.

Yeah, but what does prewarm.sh do?

It actually waits for stuff to install before shutting down tomcat.

Same as

LUCEE_ENABLE_WARMUP=true /opt/lucee/startup.sh

Okay, if not using the official docket image, you can export the LUCEE-ENABLE_WARMUP var instead.

Just call start.sh or catalina.run with

LUCEE_ENABLE_WARMUP=true

Export that var instead to make it reliable

My runtime deploy.log (which is actually being putputted to the console) has:

INFO [main] org.apache.catalina.startup.HostConfig.deployDirectory Deployment of web application directory [/opt/lucee/webapps/ROOT] has finished in [1,253] ms

Sorry, I didn't read the whole thread so I might be missing the angle here. However, @Adam Cameron, please do note my conclusion about proper extension installation here. It's got the link to the change in the official image, too: https://cfml.slack.com/archives/C06TA0A9W/p1691615352585169?thread_ts=1691140244.277319&cid=C06TA0A9W

What version of lucee? so I can check

What version of lucee? so I can check

if that was for me: sorry, i'm not sure how to cross-check that easily.

@jamiejackson not for you. Trying to replicate the issue that @Adam Cameron is having so I did a bunch of tests.

We should prolly split Jamie's thread and Adam's. While they're both related to extensions, I I'm not sure they're both having the same issue.

That would be good, yes.

What version of lucee are you using so I can double confirm the error pls/

5.3.9.173

@Adam Cameron -- Using the official images, I've been successful at doing an "ENV LUCEE_EXTENSIONS ......." in my Dockerfile, copy in my code, then run the /usr/local/tomcat/bin/prewarm.sh That prewarm does download all the extensions and any dependencies they may have into the image layer. I have a few hundred sites running it in that manner on both Docker, k8 and Lightsail Containers.

which json functions?