On Fri, Jun 16, 2006 at 10:28:50AM +0800, Freeman Liu wrote: > > Will you treat differently between manual-unplug-failure and > hardware-malfunction-failure ? > What about to handle handle human misoperation differently from hardware > failure ? Seems generating > different events in these situation will help to achieve that.
Yes, a hotplug event does not correspond to a fault. Also, if a new device is plugged into the same physical slot, then there will be a per-pool toggle switch to enable automatic replace/resilver. If you plug the same device back in, it will automatically resilver regardless. > I'm not sure if a unplug event generates FMA event, but it will > generate a unplug event ( I think it some different from a error ) > which will be caught by tamarack ( on going project by Artem > Kachitchkin). Currently, the DE subscribes to the sysevents corresponding to hotplug events. Moving forward, we want to integrate this into fmd/libtopo so that any fmd module can react to these events without having to manually plumb up the underlying mechanism. - Eric -- Eric Schrock, Solaris Kernel Development http://blogs.sun.com/eschrock _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
