In vplat_teardown, part of the teardown is: * Call zone_shutdown, which will run the ZSD shutdown hooks * Tear down network interfaces * Tear down TCP connections * Unmount filesystems. * Call zone_destroy, which will run the ZSD destroy hooks
Why is this OK for NFS over TCP, in particular NFSv4? We've called the ZSD shutdown hook, which told NFS to stop a lot of kthreads; we've torn down the network interface which might have been the one it was using; we've aborted its TCP connection; and now we're asking it to unmount. Does this work only because we're trying a forced unmount first in unmount_filesystems? Isn't that somewhat rude? I ask because I've attempted to create something like this behavior in a somewhat different environment (soft rebooting -- the global zone) and am ending up with (only some) NFS mounts unaccessable. Thanks much. --nwf; This message posted from opensolaris.org _______________________________________________ zones-discuss mailing list zones-discuss@opensolaris.org
