Samba broken again - any ideas?

[Luke Kenneth Casson Leighton]

hi graham,

what size are your log files at, in the var directory?  there was a 
problem with wrap-around on the log files, which we discovered recently 
(unfortunately, the error message itself was being truncated :-) :-)

lukes


On Fri, 7 Nov 1997, Graham Leggett wrote:

> Hi all,
> 
> I have a samba installation v1.9.17p4 on a Solaris 2.5.1 box on a
> different subnet to our network.
> 
> A short while ago, the box grew extremely loaded (we don't know why, but
> we don't suspect samba), and people (Win95) were no longer able to
> browse shares, or even telnet in.
> 
> I restarted the Solaris box in the hope things would settle down, but
> still noone (Win95) could browse shares. I upgraded samba to
> v1.9.18alpha9. Still noone could browse shares. I reverted back to
> v1.9.17p4, still none could browse shares.
> 
> - In all cases, both smbd and nmbd processes were killed.
> - In all cases, smbclient from a linux box on our network successfully
> browsed shares 100%.
> 
> Anyone know what on earth is going on?
> 
> A general question: Why is samba so unreliable? I have used different
> versions of samba for over two years, and between then and now samba has
> been very unreliable, and doesn't look at becoming more reliable now.

hm.  interesting.  has anyone else had similar reliability problems with 
Solaris 2.5.1?  have you got any patches applied?

how much memory do you have in your server, graham?

> BUT - having said this I have almost never had the slightest problem
> with samba clients connecting to samba shares, all my problems occur
> Win95 -> samba. BUT having said THAT, Win95 -> Win95 works fine.
> 
> Just what standard does samba use? The same as microsoft's?

well, the original project was when andrew packet-traced an SMB server (i 
think it was a DEC pathworks server) without even knowing what SMB was, 
several years ago.

the X-Open group published an SMB standard (in 1983, i think).  IBM
implemented this standard to the letter; microsoft made their own
"improvements".

microsoft kept their development of SMB quiet for about ten to fifteen 
years.  in august 96, they launched the CIFS initiative.  just a few 
weeks after Sun announced WebNFS.

so, all of us who have CIFS clients and servers - microsoft included - have
had to deal with the increase in understanding of the CIFS protocol, and
attempt to maintain compatibility with other CIFS vendors (hence, the CIFS
workshops).  the samba team's experience at reverse-engineering this
protocol, almost exclusively from packet traces, has actually *helped*
microsoft to write up their CIFS documentation. 

that's not to say, that with a document running at around 250 pages, with 
supplementary documentation at approximately another 100 pages, that 
every implementation's going to be perfect...


luke


<a href="mailto:lkcl at switchboard.net"  > Luke Kenneth Casson Leighton </a>
<a href="http://mailhost.cb1.com/~lkcl"> Lynx2.7-friendly Home Page   </a>
<br><b> "Apply the Laws of Nature to your environment because your
         environment applies the Laws of Nature to you"               </b>