Lucee 6.1.1.118 not stable

dev support
1

After updating to 6.1.1.118 Lucee server goes down at least 1 time a day and I have to restart it.
I downgraded to 6.1.0.243 and no problems for over 2 weeks. I upgraded again to 6.1.1.118 and after a day and a half it was Down again.
Has anyone else had stability issues with 6.1.1.118?

OS: CentOS Linux 7.9.2009 (Linux (3.10.0-1062.18.1.el7.x86_64) 64bit)
Java Version: 11.0.6 (AdoptOpenJDK) 64bit
Tomcat Version: Apache Tomcat/9.0.31
Lucee Version: 6.1.1.118

2

Just out of curiosity, if you want reply: why did you upgrade to version 6?
Version 5 seems much more reliable to me at the moment.

Thanks :blush:

3

Many issues here.

Centos 7.9 is end of life. You need to upgrade this ASAP

Tomcat 9.0.31 is overly outdated. Latest version of Tomcat 9 is 9.0.98
–Upgrade it

try that first, as I have Lucee 6.X in production across multiple servers without issue.

1 Like
4

6.1.0.243 works great
6.1.1.118 crashes

1 Like
5

I’m thinking of redoing a new plesk server with Almalinux 9.
I need to check what the best configuration is for Lucee.

6

Alamalinux 9 is what I use for my default Linux server install of it.

As for lucee, really comes down to what you are trying to do and are technically comfortable with.

You can run Lucee on FreeBSD, AIX or anything else that will run Tomcat.

7

Michele, I gather that your primary goal here is simply to find if others might know of an instability issue in that specific Lucee version, is that right?

And if no one confirms it, are you then interested in diagnosing the problem? Or will you just camp at the older one and wait to see if things get sorted out?

It’s a sincere question, without judgment.

I’m just saying there’s much more we could ask and that you could offer, to better understand and perhaps resolve why your “server goes down”. There’s often far more specific aspects to that, which could identify the specific problem.

It might even serve to help the support folks address an issue in Lucee or a related library. But for you we might at least identify a possible workaround…and one that would allow you to update and continue to update Lucee until some Lucee issue is resolved.

But I realize some won’t want to “bother” with that sort of effort, or may feel it’s not their job, or again are simply ok to park at an older version and wait. I’m just trying to tease out where things stand for you, as well as to guide us in next steps if no known instability issue for that version is identified.

8

I would be happy to help you, but I need you to tell me how.
The problem is on the machine in production, and I can’t do experiments

9
  1. @carehart is by far the expert, on all things ColdFusion servers – period.
  2. If its a paid production site, it maybe well worth having him poke around your setup, as his one hour of suggestions is literally better than a dozen hours by senior engineers all of whom deal with tomcat, ColdFusion and other application servers.
  3. Having only one server that’s always in production, not a good idea long term. I suggest you snap shot it / clone it, do what ever you must to have a working secondary replica of it. Then build a non-production server for testing and or familiarize yourself with redundancy.
1 Like
10

Terry, thanks for the kind (if overstated) regards. :slight_smile:

Michele, the goal of my question was to find simply if you DID want to dig into diagnosing things. Now that you’ve confirmed you do, I and others can offer more thoughts. I didn’t then, for the reasons stated.

And now you add “this is production”, and “can’t do experiments”. Sounds like you’ve been burned before. Diagnostics need not entail impact on prod.

So first, please clarify what “server goes down” has meant in this case? Lucee stopped responding? Even the admin? And I presume you looked for but found “nothing in the logs” at the time?

And is it that you have nothing in the form of any monitoring of Lucee? No fr, no seefusion, no apm, no jdk tools? And by “no experiments”, does that mean you won’t consider enabling them, even to possibly diagnose and resolve this problem (and other future ones).

Again, you may just want the solution, and someone else may soon gather enough from what’s been said so far to offer one. I’m asking more questions to draw out what you do and don’t know, and what we may or may not be able to obtain through this dialogue.

1 Like
11

Yeah, I’d be curious what the last operation is before the crashes. Is it something weird like PDF creation? Accessing a drive path that doesn’t exist? Just making stuff up, but meant to illustrate that there’s “regular website / CF stuff” and there’s more complex features that may have a problem on the specific hardware, OS, etc.

2 Likes
12

For us it’s the scheduled tasks that are killing the server - the ticket is here.

1 Like
13

SAME issue here. Server has been running fine for over a year. Then saw an update in the Admin last week so I updated, and now server goes DOWN every couple of days. Rolling back to 6.1.0.243.

14

Same here as well, and same with 6.2, as far up as 6.2.0.298-RC. It’s always java heap space for us in catalina.out, but haven’t figured out why yet. Downgrading also worked as a bandaid though we’re trying to narrow down further on a staging server without luck yet.

15

Same here. Every build of 6.x has caused similar issues for us. Narrowed it down to using threads and database connections in them and/or scheduled tasks using threads.

1 Like
16

Same problems here - our case is perhaps a bit more esoteric, as we’re doing object serialization at the Java level, but the bottom line for us is that 6.1.1.118 is completely unusable for our purposes.

Our specific reason is that the work-arounds we built with Lucee 5.x for LDEV-2138 don’t work, and our attempts to come up with a new work-around are stymied by a class loader problem (6.1.118 says ‘no such method exists’ for a method we can call successfully with 6.1.0.243). We still have to probe around and find what we’re looking for, but we can’t even start if we can’t get Lucee to call our Java methods

We still have to probe around and find a new work-around for LDEV-2138, but at least with 6.1.0.243 we can reliably call our Java methods.

=====
Java 11 update 25
Lucee 6.x running on Alma Linux 9 (Intel)

17

There is nothing overly complex on in my code, it’s quite basic. I was a little concerning to see these threads as I am on 6.1.1.118 , however I can confirm that I have had no problems at all.

I can’t remember when I upgraded but it’s been a while so likely shortly after it was released which I believe was November.

Tech stack

  • Lucee 6.1.1.118
  • Java 11.0.21 (Eclipse Adoptium) 64bit
  • Windows Server 2022 (10.0) 64bit
  • IIS 10
18

On 6.2 the problem seems to be solved.
I updated to 6.2 already 10 days ago and since then the server has not crashed.
Instead every time I tried to update to 6.1.1.118 after about 1 day it was down. In the logs I only found a timeout without specific indications.

2 Likes