December 5, 2017 at 8:34 am #544506
Situation: Tomcat is started in MyEclipse. I click on the “Terminate” button in the console to shut it down, as I need to restart many times throughout the day and have no intention of waiting on a graceful shutdown.
Apparently it’s possible for MyEclipse to get confused and think that the Tomcat server is still running. When trying to start the server, it calls the stop routine which fails to connect to the server.
Instead of recovering from this, all you can do is end the “Stopping the Tomcat instance” task in the Progress view. Tomcat will then display the server as running, even though it is most certainly not…
Occurs in CI 9. Does not occur in CI 8.December 5, 2017 at 1:27 pm #544539
It’s odd that this occurs only in CI 9 – I’m not aware of changes that should have caused a regression here; we’ll take a look, of course.
1) I assume the difference you are seeing, was that in both CI 8 and CI 9, pressing the terminate button does terminate the running Tomcat instance – but in CI 9, the stopping routine doesn’t correctly complete, while it does in CI 8. Does this sound right?
2) Do you run into this problem all the time – no Terminate works as expected? Or does it only occur sometimes?
3) What version of Tomcat are you connecting to? I’m assuming it’s an external Tomcat instance and not the embedded MyEclipse Tomcat?December 6, 2017 at 2:15 am #544553
1) That’s right. I terminated the Tomcat process by using the “Terminate” button in the Console view. I noticed that Tomcat was still marked as running (“debugging” status) in the Servers view. I could not stop the server afterward without restarting MyEclipse.
2) I don’t know. After I saw that & the scrollbar issue I quickly reverted back to CI 8.
3) It’s an external Tomcat instance running in “c:\development\Tomcat\apache-tomcat-7.0.59”. The configuration has been modified in a way that MyEclipse does not overwrite our configuration files (i.e., “c:\development\Tomcat\apache-tomcat-7.0.59\conf” is not controlled by MyEclipse). So all we’re doing is starting and stopping a Tomcat instance, there is no management other than attaching the debugger to the Tomcat instance. It has worked (and continues to work) just fine in CI 8.
However, it would not surprise me if this issue also occurs in CI 8. The complexity of (My)Eclipse is so great that it’s often impossible to tell exactly what it’s doing.
For example, yesterday, a colleague was trying to open a Java file. The editor screen would claim that the file did not exist while both the file system (checking w/ Windows Explorer on a NTFS file system) and the Package Explorer (file was visible and continued to be visible even when refreshing the project) did not concur. A restart of MyEclipse helped then as well. But I wouldn’t know what caused it…