(cc'ing cifs-discuss since I think this is related) On Fri, Jun 5, 2009 at 17:24, Justin Lee Ewing <jlewing at jrleindustries.com> wrote: > I upgraded one of our local test servers here from 2008.11 to 2009.06 and am > now seeing strange behavior during an reboot or shutdown process (using init > 6 or init 5). (...) > ? root ?1860 ? ? 7 ?23 12:02:47 ? ? ? ? ? ? 0:51 /usr/sbin/sharemgr stop zfs > > in the process table... the problem is that the command never completes and > the system never reboots or shuts down. > > The server is sharing out a large ZFS filesystem using Samba. > > Is anyone else seeing this? This could be the same as the CIFS issue I'm seeing [1]. Try "ps -ef | grep smbd", then kill -9 that process, and then try ps again. If nothing changes, it's almost certainly the same issue. "svcs -v smb/server" is also useful; for me it shows that its state is "online" and nstate is "offline", but it never makes it all the way to shutting off. Eric Mudama also experienced these symptoms, although for him the effect was softened with 2009.06 aka build 111b.
I haven't found a solution to the problem. Hard rebooting (i.e., "reboot" rather than "init 6") causes CIFS shares to work until a client is rebooted, but then smbd goes back to its unhappy mode. Will [1]: http://www.mail-archive.com/cifs-discuss at opensolaris.org/msg01024.html
