Menu
▾
▴
[Wrapper-user] Re: Converting RMID into a NT/2000 service - Windows Crash
|
From: <zge...@ex...> - 2004-06-04 17:34:47
|
I now have several services that use JSW on Win 2000. Everything works fine except that ever since I started using JSW - my machine seems to crash for no apparent reason. I am still in dev. mode - hence I frequently install and remove and re-install the services from the machine WITHOUT REBOOTING after each "remove". Is this an issue? The reason I ask is that if I do reboot after each remove the crashes do not seem to happen as often. Of course, this may have nothing to do with JSW and everything to do with my machine. But has anyone else seen this behavior? Could it have something to do with the fact that I do NOT reboot each time I remove a set of services. Thanks in advance. Zac George > > Subject: > Re: [Wrapper-user] Converting RMID into a NT/2000 service > From: > ""<zge...@ex...> > Date: > Fri, 28 May 2004 12:26:40 -0400 > To: > wra...@li... > > > Leif, > > Thanks for the response. My first attempt at this was to try the > > approach you recommend below > > (in fact that's the approach I took with the java > > Name Service which worked). > > I did not have much success with it for rmid. > > > > However, I did NOT also set the policy file to give specific > > permissions to the wrapper jar. Wrapper jar was bundled into my > > codebase jar file. I'll try the route you suggest once again. > > > > The second issue with the approach you suggest was that I could not > > find a way to pass the "-C" command line options that > > I can typically pass on the command line to rmid.exe for > > passing on to child JVMs if I call on the sun implemenation > > directly. > > > > Even if there is a way around the above two issues, > > I believe there may still be a third issue - the rmid daemon will > > activate JVMs for activatable objects. I need those JVMs to survive > > login/logouts. The launch sequence is as follows: > > wrapper.exe->rmid.exe->java.exe (contains activatable object). Since > > wrapper does not own these "sub" processes they clearly will not > > survive the login/logout. What would have been ideal was for a > > parent process to own the child process. THe child process may choose > > to exit on its own - if it did not make that choice, then the parent > > process must control its life cycle. > > > > Again, thanks for your response. > > -- > > Zac > > _______________________________________________ Join Excite! - http://www.excite.com The most personalized portal on the Web! |
