This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
pcm ran into trouble on Fedora Core 3. More details to follow from him.
I've seen the same behavior sometimes with TPR3 builds on different machines with different OS. In fact, server gets started and I was able to update its status in the IDE using Refresh command. I have no idea what are the steps to 100% reproduce this issue.
How long does it take to start the server? MAybe longer than 2 or 3 minutes? (possible with OpenESB?) What apps are deployed? (attach the domain.xml) In this case, Nb will timeout. YOu can alway to a node refresh to check again the server status. See the log file entries for timestamps... Or attach the log, as requested on the previous bug. Vince: I do not think it is a good idea to close a bug thread and reopen a different one: where are all the questions I asked? Lost? Also, this is really a P1 for this customer on these machines... The fact you cannot reproduce it does not decrease the priority.
PCM: Please provide detailed responses to Ludo's question. PCM: obvioulsy more info would help...I am not sure why you do not provide it:-( What is the reproducible case? What is the ratio of failure in %? What is the state of the system after (ie server is running or not?) What happens if you restart the IDE: does it see the server running? What is special with the network settings on this machine? (can we remote login to analyse? If yes, provide the machine name, ip) How many CPU? Which JDK? When this bug was opened las year, was it on the same machine? Do you have another machine to try to reproduce?
from phone conversation with PCM steps to reproduce initial state.. NetBeans and server are RUNNING in the IDE: 1. Use Stop action to stop the instance instance stops after some amount of time 2. Use Start action to start the same instance startup appears to fail. (get the failed to start message) 3. use command-line `kill` to get rid of the app server instance's process 4. Use the start action this time it is successful. PCM: step 1 and 2 may be replaces by single use of Restart action?
Vince, This steps are correct.
Does this instance contains openEJB, IM, JBI modules? on my system, a clean GalssFish instance starts in 15 seconds, but a clean instance from the EE SDK tool bundle takes 62 seconds... Quadrupling the start time in the EE SDK tool bundle in not acceptable imo. Obvioulsy, the timeouts in NetBeans are not big enough. We can increase them, but the point is that the real fix is to work on the instance startup time.
Also, why do you kill the process? Does a Server nod "Refresh" popup menu works? If yes, then it is a P3: just select this refresh menu so that Nb sees the running server.
As for me, Refresh always helps, so I never use kill command.
PCM: I am not pleased with the complete lack of extra info on this bug. Obviously this is not a show stopper since you are not looking as being stuck with this bug:-) Worakorund seems to refresh the server node, so it is a P3. We'll increase the timeout value on the pluign side (even maybe diaply a dialog to ask if the ide needs to continue waiting for a complete startup). If you could attach the server log file, maybe give use the machine name/ip so that we can try, it would still be valuable.
This is not consistantly reproducible. I am OK to make this as P3 for now. If this reproduced again. I will add more details.
I increase the time out from the IDE to 5 minutes to see if the server is running. Closing, but open a new bug if this happens again, with detailled info, and log files.