<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> <blockquote cite="mid:FAC6DFEF4E4DDF41A31D7B7524D6DF87040653F6@bbcxues14.national.core.bbc.co.uk" type="cite"> <pre wrap=""> That's surprising, but good news. I hope you don't rely on it </pre> </blockquote> Oh, we do, after running it for half a year without problems.... the hard part was to get the standard check scripts running as they won't compile under cygwin. The main nagios executable compiles and runs nicely (hem... almost..) with --enable-cygwin . <blockquote cite="mid:FAC6DFEF4E4DDF41A31D7B7524D6DF87040653F6@bbcxues14.national.core.bbc.co.uk" type="cite"> <pre wrap=""> Interesting definition of "quite successfully". </pre> </blockquote> Well, as said, the error did not appear always. It happens now and then, sometimes twice a day and then the system runs for weeks. <blockquote cite="mid:FAC6DFEF4E4DDF41A31D7B7524D6DF87040653F6@bbcxues14.national.core.bbc.co.uk" type="cite"> <pre wrap="">I've never seen anything like it. </pre> </blockquote> I haven't and we are running other nagios installations too, on several different linux installations. <blockquote cite="mid:FAC6DFEF4E4DDF41A31D7B7524D6DF87040653F6@bbcxues14.national.core.bbc.co.uk" type="cite"> <pre wrap=""> Cygwin != unix. At the core, windows deals with files and sockets in a very different way to linux and solaris (and probably other unixes) It wouldn't surprise me if the something changed with the status file. If you can't install a linux distro onto the box, run up a virtualbox environment to host nagios in. It wouldn't be perfect, but will probably be better than cygwin. </pre> </blockquote> For reasons that are quite complicated in this case it is not an option to run nagios in a virtual box. I posed this question here mainly to see if this happens with unices also - if not, it is probably something for the cygwin mailing list. I just observed that after ending one of the two existing nagios processes with the windows task manager, things continued to run normally without a restart. So now I suspect it is either a blocked file access or some fork() problem- these sometimes happen with cygwin. I enabled the large installation quirks that should reduce the number of forks, let's see if that helps. Regards, Jakob </body> </html>