Hello, I think (in fact I'm quite sure) I have found the following bug in Swat. In my configuration I use the global option "invalid users=@class". When I set this option using SWAT, all works fine after "commit changes". Users who belong to the "class" group cannot access the server. But when I erase this value using SWAT and do "commit changes", everything SEEMS to be ok, but in fact these users still can't access the server. Only when I restart the smb daemon (using SWAT, Status) everything is really ok and users of the "class" group can access the server. But I do not want to restart the smb daemon since this breaks all existing connections. Does anybody has any idea? Can anyone of the Samba team confirm this issue or not? Thanks alot, Werner Maes KULeuven
But why does it sometimes work when you do "commit changes". For example, when you change invalid users from empty to @class and do "commit changes", you don't have to restart samba. But when you erase this value, you need to restart samba. That's not very logic in my opinion. Werner KYT wrote:> I am not from the Samba team, however, If I am not wrong, when you "commit > changes" you have commit to change the configuration files (smb.conf). You > will still need to restart samba to APPLY the changes, this is where > restarting of smb daemon comes in. :-) > > Don't think this is classified as a bug. (Normally I will do a > samba -restart) > > Werner Maes wrote in message <398EE12C.68C66EC5@cc.kuleuven.ac.be>... > >Hello, > > > >I think (in fact I'm quite sure) I have found the following bug in Swat. > > > >In my configuration I use the global option "invalid users=@class". > >When I set this option using SWAT, all works fine after "commit > >changes". > >Users who belong to the "class" group cannot access the server. > > > >But when I erase this value using SWAT and do "commit changes", > >everything > > ---snipped---
I have done some new tests and I can now confirm that waiting more than one minute doesn't work, nor doing a reload of the smb.conf file through "smb reload". The only thing that works is restarting the smb daemon. Werner Maes KULeuven Mikkel K Johnsen wrote:> Yes, I have read that too, but I have not tested it. (60 sec). Doesn't it > work after 60 sec ? > > You can run "smb reload", that makes samba reread the smb.conf file. (or > send it HUP) > > -- Mikkel > > "Werner Maes" <werner.maes@cc.kuleuven.ac.be> wrote in message > news:398EE12C.68C66EC5@cc.kuleuven.ac.be... > > Hello, > > > > I think (in fact I'm quite sure) I have found the following bug in Swat. > > > > In my configuration I use the global option "invalid users=@class". > > When I set this option using SWAT, all works fine after "commit > > changes". > > Users who belong to the "class" group cannot access the server. > > > > But when I erase this value using SWAT and do "commit changes", > > everything > > SEEMS to be ok, but in fact these users still can't access the server. > > Only when I restart the smb daemon (using SWAT, Status) everything is > > really ok and users of the "class" group can access the server. But I do > > not want to restart the smb daemon since this breaks all existing > > connections. > > > > Does anybody has any idea? > > Can anyone of the Samba team confirm this issue or not? > > > > Thanks alot, > > > > Werner Maes > > KULeuven > > > >
What I understand of the commit changes in SWAT is to save the changes in smb.conf. But for the changes to take effect, u have to restart the services. Right, Dave?!? Mon Hello, I think (in fact I'm quite sure) I have found the following bug in Swat. In my configuration I use the global option "invalid users=@class". When I set this option using SWAT, all works fine after "commit changes". Users who belong to the "class" group cannot access the server. But when I erase this value using SWAT and do "commit changes", everything SEEMS to be ok, but in fact these users still can't access the server. Only when I restart the smb daemon (using SWAT, Status) everything is really ok and users of the "class" group can access the server. But I do not want to restart the smb daemon since this breaks all existing connections. Does anybody has any idea? Can anyone of the Samba team confirm this issue or not? Thanks alot, Werner Maes KULeuven