Re: mystery deaths with deployed app
Re: mystery deaths with deployed app
- Subject: Re: mystery deaths with deployed app
- From: Timothy Worman <email@hidden>
- Date: Tue, 07 Apr 2015 18:49:37 -0700
The JVM is /Library/Java/JavaVirtualMachines/jdk1.8.0_25.jdk on Snow Leopard. I have these two beauties waiting in the system.log.
Apr 7 12:40:11 apps java[3715]: *** NSInvocation: warning: object 0x11a108400 of class 'ThreadUtilities' does not implement methodSignatureForSelector: -- trouble ahead
Apr 7 12:40:11 apps java[3715]: *** NSInvocation: warning: object 0x11a108400 of class 'ThreadUtilities' does not implement doesNotRecognizeSelector: -- abort
Apr 7 12:40:12 apps com.apple.ReportCrash.Root[4121]: 2015-04-07 12:40:12.444 ReportCrash[4121:2a03] Saved crash report for java[3715] version 1.0 (1.0) to /Library/Logs/DiagnosticReports/java_2015-04-07-124012_localhost.crash
Apr 7 12:55:58 apps java[4152]: *** NSInvocation: warning: object 0x1186a8400 of class 'ThreadUtilities' does not implement methodSignatureForSelector: -- trouble ahead
Apr 7 12:55:58 apps java[4152]: *** NSInvocation: warning: object 0x1186a8400 of class 'ThreadUtilities' does not implement doesNotRecognizeSelector: -- abort
Apr 7 12:55:58 apps com.apple.ReportCrash.Root[4364]: 2015-04-07 12:55:58.861 ReportCrash[4364:2a03] Saved crash report for java[4152] version 1.0 (1.0) to /Library/Logs/DiagnosticReports/java_2015-04-07-125558_localhost.crash
I’ll have to look at those crash logs more carefully. Maybe there’s a change in Java 8 that causes an incompatibility here.
I’m not using any instance rescheduling. Nothing calling exit(). Maybe Application.terminate() isn’t even getting called? There isn’t any indication in the app’s logs that it terminated.
Tim Worman
UCLA GSE&IS
> On Apr 7, 2015, at 1:46 PM, Chuck Hill <email@hidden> wrote:
>
> The JVM just ending with no message is odd. Some things to consider:
> - look at the system log, you might have developed a hardware problem that is causing the process to seg fault (or similar) and exit
> - do you have scheduling on and something is preventing it from restarting?
> - do you have some code that is calling exit() on the Runtime?
> - Wonder has some code that will kill an instance if it runs out of memory (and maybe if requests take too long, I don’t recall offhand what got left in), but I would expect that to log something
> - override terminate() in Application and log out a stack trace
> - did you change to a different JVM? Maybe you got one with a defect?
>
>
> Chuck
>
>
> On 2015-04-07, 1:41 PM, "Timothy Worman" wrote:
>
> I have an app that has been in service for a long time in various forms. Recently I deployed an updated version that has been in development in recent months. The update has been experiencing “silent” deaths every once in a while. There is nothing in the app log that indicates any exception that may be occurring when this happens. What I do see is loads of messages in the logs re: CookieParser: Found a null cookie value in….
>
> However, this was occurring before. Sometimes I’ll simply see a few logins that appear normal, a bunch of CookieParser errors, then the app log just ends. Maybe an exception is getting swallowed somewhere but I can’t find anything like that.
>
> Any suggestions?
>
> Tim Worman
> UCLA GSE&IS
>
>
>
>
> _______________________________________________
> Do not post admin requests to the list. They will be ignored.
> Webobjects-dev mailing list (email@hidden)
> Help/Unsubscribe/Update your Subscription:
>
> This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden