Hello all. We have CF10 Enterprise on a windows server 2008 R2 server. IIS 7.5. Patched through Update 16. We've had a site running fine for a while now and it was time to add another application. I verified I wouldn't run out of memory and created a new instance. The instance was created fine but it would not start when I clicked the start button. I have Fusion Reactor installed on the server and it showed activity from the start click-then the activity stopped. I had selected the Create Windows service check box so a service got created. I tried starting just the service and the service wouldn't start either. Meanwhile the original site continued to run.
I checked to make sure the the ports used for the new instance were not already allocated, and they were not. I checked permissions-no issue there. There is plenty of RAM and disk space.
I tried removing the instance and creating it without the Windows service - no change in behavior.
The CF logs do not show a failure. The server's Event viewer log tells me the application (instance name) does not exist but the instance is showing CF admin.
I am really stuck here and I have a public facing application due to go out very soon.
Thanks for your time.
Andy, I have a few thoughts for you to consider, since it could be any number of things
First, can you confirm that you only see one ColdFusion.exe process listed in Task Manger when you say only one is running? That’s key, to confirm if Windows thinks the instance is running, even if other means you use would suggest otherwise.
Second, did you look specifically at the ColdFusion-error and ColdFusion-out logs (in the instance’s logs folder)? And if only one ColdFusion.exe is running, did you look at the log entries right around the time you tried to start the other, to see if there was info there? And if there is more than one ColdFusion.exe running, can you stop the one that is the troublesome one, and THEN view what those logs may say at the bottom, as soon as you stop the process, and then when you try to start it? There may be info to explain things.
Third, if the process immediately fails to start, it sounds on first blush like a JVM problem, where the jvm.config of the instance you’re trying to run is pointing to a JVM that doesn’t work. By default, I’m pretty sure it should be pointing at the same JVM that the cfusion instance points to (when the new instance is first created, at least). Can you look in the jvm.config for both instances (in each instance’s bin folder), and look at the first uncommented line, which will show the java.home value. What are they, for each instance?
And can you confirm that the working instance has not been restarted since perhaps any change was made to that value (in the working instance)? It’s possible that a change was made to that instance’s java home location, but the instance has not been restarted since then. But if the new instance inherited those settings, I would be trying them as soon as you tried to start that new instance—and if you tried to restart the other instance, if it’s not been started since it was changed.
Fourth, if it’s not this, my next question (if the logs really show nothing, using the approach I showed above) would be to propose that you look for logs instead in that same bin folder where the jvm.config lives, within the new instance’s folders. Yes, that bin folder can hold logs, if the JVM crashes with an hs_err_pid*.log file being created. If there are none, then for now I’m out of ideas.
But let’s see what you find regarding these four possibilities above.
Hi Charlie. Thanks for responding.
1. There is only one ColdFusion.exe running and one ColdFusionsvc.exe running.
2. Neither log shows any activity or error relating to my attempt to start the instance.
3.There is a 5 second pause before the process stops.I took a look at the JVM arguments and found something strange. There was a reference to a FusionReactor instance that didn't exist. It looks like if you have FR and you create an instance, FR tucks in this reference. I took out the reference and the new CF instance fired up. Now will be the trick of configuring FR to watch over the new instance and see if everything keeps working.
I'll need an evening to finish configuring the server but thanks to your help, I think I found the issue. ( I will know in a couple of days for sure when I can get access to the server again)