On Wed, 2005-01-12 at 09:57 -0500, goggin, edward wrote: > Shouldn't signals be masked while calling suspending a multipath mapped > device in > order to reload its map in multipath/main.c:domap()? > Having removed the suspend / resume cycle around table reload, I guess the process can be killed without unwanted consequences. Thanks for reporting, -- christophe varoqui <christophe.varoqui@xxxxxxx>