Before doing much with Raspberry Pi devices, I had a reliable network where I could share files between Linux machines and Windows boxes with ease. Windows to Windows and Linux to Linux also worked great. Samba and Samba clients mostly just worked out of the box. Even changed my workgroup name on a few machines, and lo and behold, the Samba Client program would see the new workgroup and let me access machines on it.
I had played with a couple of Pi's before even trying PiStar, just to be learning things. Never did much more than a few applications, and never had them on the network long enough to cause problems. But once I got three instances of PiStar running, things began to change. Shares began to disappear. Whole machines and workgroups would become invisible. I did numerous things to try and resolve the issues, but made things worse in some cases. I also have three extra Pi boards, and some of them were on the network at this time as well. All were running Raspbian Stretch. On only one of these did I install the full Samba package, and that was after the problems started.
So over the last couple of days, I shut off every Pi that was on my network, and went about restoring all the settings that I had changed on various systems. Even had to completely re-install and re-configure Samba on one Linux Mint system. But after doing this, my network was mostly back to normal.
So I offer this as a cautionary tale to anyone else that has a Samba sharing network already in place. Putting multiple PiStar hot spots on your existing network may break some things. It sure did for me.
Doug
K9DJN
Too many Pi's on a network spoils the Samba soup
Re: Too many Pi's on a network spoils the Samba soup
Did you make sure you changed the RPi hostnames so that each system had a unique name?
--
AF6VN
Dennis L Bieber
Re: Too many Pi's on a network spoils the Samba soup
Oh, absolutely. pi-star0, pi-star1 and pi-star2 are the hostnames. I could reach them by name from any Linux system at the time I was having the problem, but not the Windows system. I'm planning on re-powering one at a time and then making sure I can reach it from both Linux and Windows. It's probably only one that's causing the issue.
Re: Too many Pi's on a network spoils the Samba soup
Never had these problems. Here's my Windows 10 network:
And this is how I install Samba on my PI's:
$ sudo apt-get install samba samba-common-bin
$ sudo nano /etc/samba/smb.conf
workgroup = WORKGROUP
wins support = yes
Add a section at the bottom:
[pi]
comment=PiHome
path=/
browseable=Yes
writeable=Yes
only guest=no
create mask=0777
directory mask=0777
public=yes
$ sudo smbpasswd -a pi-star
(now enter password twice: raspberry)
$ sudo /etc/init.d/samba restart
But thanks for the nice story !
73's
Adri
And this is how I install Samba on my PI's:
$ sudo apt-get install samba samba-common-bin
$ sudo nano /etc/samba/smb.conf
workgroup = WORKGROUP
wins support = yes
Add a section at the bottom:
[pi]
comment=PiHome
path=/
browseable=Yes
writeable=Yes
only guest=no
create mask=0777
directory mask=0777
public=yes
$ sudo smbpasswd -a pi-star
(now enter password twice: raspberry)
$ sudo /etc/init.d/samba restart
But thanks for the nice story !
73's
Adri
Re: Too many Pi's on a network spoils the Samba soup
Ha, well it hasn't been such a nice story for me. I can now see a network picture like you posted, but when all the Pi's were on, the whole workgroup just disappeared. It's going to be a slow and arduous process to see which Pi is the problem child, but I WILL find him!
And thankfully, a cold weekend is coming up, so there will be plenty of time to test things.
Cheers,
Doug
K9DJN
And thankfully, a cold weekend is coming up, so there will be plenty of time to test things.
Cheers,
Doug
K9DJN
Re: Too many Pi's on a network spoils the Samba soup
He Doug, sounds like your network has some issues with the Master Browser elections.
Read this article, perhaps it speeds up your search for the guilty PI...
https://scottiestech.info/2009/02/14/ho ... orkgroup/
These are the results for my network. I t seems my RPI-DV4MINI is the Master Browser!
In your case that should be the suspected bad boy...
Good luck!
Adri
Read this article, perhaps it speeds up your search for the guilty PI...
https://scottiestech.info/2009/02/14/ho ... orkgroup/
These are the results for my network. I t seems my RPI-DV4MINI is the Master Browser!
In your case that should be the suspected bad boy...
Good luck!
Adri
Re: Too many Pi's on a network spoils the Samba soup
Thanks. I ran across that program the other day when I was researching this problem. Wish there was something on Linux that also revealed the Master Browser, but I do have a couple of Windows 7 boxes on my network. Will give this a whirl and see what it comes up with.
Re: Too many Pi's on a network spoils the Samba soup
There is actualy.
Above image is a SSH session to one of my PI's, showing the command nmblookup -T -M -- - and the
reply RPI-DV4MINI, 192.168.178.40 __MSBROWSE__<01>
Hope this helps....
Adri
Above image is a SSH session to one of my PI's, showing the command nmblookup -T -M -- - and the
reply RPI-DV4MINI, 192.168.178.40 __MSBROWSE__<01>
Hope this helps....
Adri
Re: Too many Pi's on a network spoils the Samba soup
Excellent. The default Samba setting is this:
server role = standalone server
Does that try to make every instance of Samba a master? I know about the election business, but in my case an old CentOS box was acting as a Preferred Master, and that may have screwed things up. Would it not be better to change all the Pi's and maybe my Mint system to "member server"?
Thanks for the help.
server role = standalone server
Does that try to make every instance of Samba a master? I know about the election business, but in my case an old CentOS box was acting as a Preferred Master, and that may have screwed things up. Would it not be better to change all the Pi's and maybe my Mint system to "member server"?
Thanks for the help.
Re: Too many Pi's on a network spoils the Samba soup
Don't think it works like that. What i understand is that "server role" controls the way samba shares behave when users login.
MasterBrowser relies on elections. As our simple WORKGROUP network does not have a dedicated domain-controller the election mechanism is needed to avoid more than one node acting as MasterBrowser.
The only way to "steer" the proces is to clear/disable samba on each node, then switch all nodes off. Next switch the one node on that you want to be the MasterBrowser and wait at least 51 minutes to settle. Then switch on the other nodes and enable Samba.
If possible do not switch-off the master browser node, as that will trigger the voting mechanism and another node will become MasterBrowser.
This is why i never experienced problems. My node "RPI-DV4MINI" is up 24/7 because it controls my main UPS backup system.
Read more: https://www.samba.org/samba/docs/4.1/ma ... onf.5.html
Adri
MasterBrowser relies on elections. As our simple WORKGROUP network does not have a dedicated domain-controller the election mechanism is needed to avoid more than one node acting as MasterBrowser.
The only way to "steer" the proces is to clear/disable samba on each node, then switch all nodes off. Next switch the one node on that you want to be the MasterBrowser and wait at least 51 minutes to settle. Then switch on the other nodes and enable Samba.
If possible do not switch-off the master browser node, as that will trigger the voting mechanism and another node will become MasterBrowser.
This is why i never experienced problems. My node "RPI-DV4MINI" is up 24/7 because it controls my main UPS backup system.
Read more: https://www.samba.org/samba/docs/4.1/ma ... onf.5.html
Adri