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.
Build: NetBeans IDE Dev (Build jdk8lambda-1778-on-20130716) VM: Java HotSpot(TM) Client VM, 21.0-b17, Java(TM) SE Runtime Environment, 1.7.0-b147 OS: SunOS User Comments: ksrini: scanning Stacktrace: java.lang.AssertionError: Resource temporarily unavailable at sun.nio.fs.AbstractPoller.cancel(AbstractPoller.java:139) at sun.nio.fs.SolarisWatchService$SolarisWatchKey.cancel(SolarisWatchService.java:189) at org.netbeans.modules.masterfs.watcher.solaris.NioNotifier.removeWatch(NioNotifier.java:84) at org.netbeans.modules.masterfs.watcher.solaris.NioNotifier.removeWatch(NioNotifier.java:63) at org.netbeans.modules.masterfs.providers.Notifier$1.removeWatch(Notifier.java:115) at org.netbeans.modules.masterfs.watcher.NotifierKeyRef.removeWatch(NotifierKeyRef.java:96)
Created attachment 137335 [details] stacktrace
The exception originates in NioNotifier. Egor, can you please check it? Thank you.
Please try to set higher opened file limit in the system: ulimit -n 10000 or even higher number. Probably filewatcher needs to handle too many files and system is not ready to that.
If a file watcher cannot be created, an IOException should be thrown, so that polling is used as a fallback. See org.netbeans.modules.masterfs.watcher.Watcher, line 244. Egor, can you please fix it in masterfs.solaris? Thank you.
I'm not sure I can: watcher was created fine, but due to lack of system resources after some time java gives assertion. And I'm sure I should not catch assertions.
(In reply to comment #5) > I'm not sure I can: watcher was created fine, but due to lack of system > resources after some time java gives assertion. Thanks, I understand now. So the exception should be caught at Watcher:234. > And I'm sure I should not catch assertions. I agree. Probably IOException should be thrown instead of AssertionError. So is it a JDK bug? What are possible solutions? Break principles and wrap AssertionErrors with text "Resource temporarily unavailable" to IOExceptions, close as WONTFIX or mark the issue as JDK bug? Any ideas? Thank you.
I would prefer marking as JDK bug, when it is fixed we can wrap the exception with a user understandable message. I tried to file jdk bug but failed, sorry, do not know how to do that now ((
I have moved the existing java incident to bug, it will appear soon on bugs. http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=8021281
(In reply to comment #8) > I have moved the existing java incident to bug, it will appear soon on bugs. > http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=8021281 Thanks! Actually I did that as well, but my bug have not appeared as well (9005347)
(In reply to comment #8) > I have moved the existing java incident to bug, it will appear soon on bugs. (In reply to comment #9) > Actually I did that as well, but my bug have not appeared as well (9005347) Thank you very much. I'm marking the issue as JDK bug and closing it for now.
This is really a Solaris 11 issue, the trouble is port_send(3C) is failing with EAGAIN because of: EAGAIN The maximum number of events per port is exceeded. The maximum allowable number of events per port is the minimum value of the process.max-port-events resource control at the time port_create(3C) was used to create the port. The max-port-events must be increased in /etc/system or using prctl(1) I think this requires documentation in NetBeans docs.
*** Bug 237411 has been marked as a duplicate of this bug. ***
*** Bug 237765 has been marked as a duplicate of this bug. ***
*** Bug 237258 has been marked as a duplicate of this bug. ***
*** Bug 239617 has been marked as a duplicate of this bug. ***