On Fri, Oct 29, 2004 at 05:43:59AM +0900, Ara.T.Howard / noaa.gov wrote:
> On Fri, 29 Oct 2004, Michael Neumann wrote:
> 
> >>
> >>Any ideas?
> >
> >Yes. By default, when you shutdown the application/webserver, all data
> >is marshalled to disk (the complete session state etc.), and next time,
> >it is load's from this file.
> >
> >This file is probably corrupted.
> 
> this shouldn't happen using pstore right?  it's transaction should work
> correctly under all conditions except nfs, power failure, etc.  or isn't 
> this
> being used?

Yes, I probably should (and will) use pstore. But corruption has never
occured on my system. Might be a Windows related problem?

Well, I don't know too much about signals unter Unix...

This is what I do:

  trap("INT") {
    trap("INT", "IGNORE")
    server.shutdown # shut WEBrick down
    self.shutdown   # shut's the application down (calls store_to_disk)
  }

Btw, does anybody knows why WEBrick needs so long to shut down? Is this
socket related?

Regards,

  Michael