I've only just noticed this slipped by without reply - sorry this took so long! On 9 February 2016 at 22:27, Michal Šmucr <msmucr@xxxxxxxxx> wrote: > > previously I've used fio mainly at Linux and OS X, so I didn't have any > issues with fio in server mode. Now I'm trying to simulate workload to > shared storage for Windows clients and this functionality with remote > control of fio instances running at Windows machines will be very handy. > Unfortunately it isn't possible to use it that way, because of missing > waitpid() syscall implementation and forking in MinGw builds for Windows. > > I've stumbled upon at some messages posted 2y ago, where Bruce explained it. > http://www.spinics.net/lists/fio/msg02565.html > > I don't know, if I got it correctly, but can be a build with Cygwin > toolchain possible workaround for this issue? > I've tried to compile current fio with Cygwin (so no MinGw toolchain) and > compiler variables override, but failed badly.. It doesn't even pass through > crc/test.c with lot of messed definitions. Is it currently possible to build > with it or MinGW is currently only way for Windows compilation of fio? Even at the time of writing fio does not build in a "full" Cygwin environment. It actually tries to be a native Windows program (hence MinGW) so it can access native Windows calls (rather than Cygwin ones) and not carry the Cygwin portability library. I suppose someone could undertake such a port but it would be nicer if server mode could be made to work with threads (but that's also new work someone would have to perform). -- Sitsofe | http://sucs.org/~sits/ -- To unsubscribe from this list: send the line "unsubscribe fio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
