From the first check I believe there is a superfluous restart of the service in

http://bugzilla.novell.com/show_bug.cgi?id=517768 User lmuelle@novell.com added comment http://bugzilla.novell.com/show_bug.cgi?id=517768#c1 Lars Müller changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO CC| |mail@cynapses.org Info Provider| |mail@cynapses.org --- Comment #1 from Lars Müller 2009-06-30 06:18:31 MDT --- the *if-down.d* case. Here I would simply stop the service in the case the interface goes down. @Andreas: DO you remind why you called restart in the *if-down.d* case? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=517768 User mail@cynapses.org added comment http://bugzilla.novell.com/show_bug.cgi?id=517768#c3 Andreas Schneider changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|mail@cynapses.org | --- Comment #3 from Andreas Schneider 2009-07-01 05:54:11 MDT --- Not really :) -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=517768 User gernot@hillier.de added comment http://bugzilla.novell.com/show_bug.cgi?id=517768#c5 --- Comment #5 from Gernot Hillier 2009-07-06 03:08:45 MDT --- Well, I'm not exactly sure if the patch from comment #2 is exactly what you want in SUSE mainline. I see a general problem in design here: if I don't got things wrong, you restart smbfs on *any* ifdown/ifup action on *any* interface independant if this interface is used for smb or not. I don't think this is a good idea. If you apply comment #2 now, an "ifdown wlan0" will shutdown your smbfs shares on eth0 which is probably not what you'd want. And as there's no output at all (rcmbfs output seems to end up in /dev/null), this is evil (tm). I think this was the reason to use "restart", not "stop" in the script to have at least functionality restored if possible. However, I think it's generally a bad idea to silently drop smbfs connections on ifup/ifdown anyways - being it temporarily or not. So my recommendation would be: if the user explicitely requests to shutdown an interface, please either let him (If I'm not mistaken, the kernel is designed in a way that you can ifdown/ifup an interface *without* remounting CIFS share using it!) without doing anything automatically - or use a clean service dependency here and just ignore the request and inform the user he is not allowed to ifdown the interface if smbfs is being used. However, as this won't be a problem for _my_ use cases (we only use an eth interface), I'll test the patch soon and let you know my results as requested. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=517768 User lmuelle@novell.com added comment http://bugzilla.novell.com/show_bug.cgi?id=517768#c7 Lars Müller changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #7 from Lars Müller 2009-07-14 16:20:59 MDT --- network:samba:STABLE and network:samba:TESTING both got the fix. You need a build-source-timestamp of at least 2175. It's possible to check this with rpm -qi samba-client -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=517768 http://bugzilla.novell.com/show_bug.cgi?id=517768#c9 Gernot Hillier changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | --- Comment #9 from Gernot Hillier 2009-11-26 15:51:58 UTC --- We found another funny issue with this which occurs even with applied patch from comment 2. When you made a manual smbfs mount on the cmdline, a DHCP renewal if lease time exceeded will kill it, because seemingly ifup is called by dhcpcd - leading to "rcsmbfs restart" a.k.a. "rcsmbfs stop; rcsmbfs start". Mounts from /etc/fstab will be available again afterwards, but no manual mounts. So I'd again like to suggest to completely removing these magic hooks as suggested in my initial comment... -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.