[Pkg-samba-maint] Bug#574283: samba: Doesn't bind to IPv4

[Martijn Uffing]

Ave people

I can reproduce this bug.  I've got no hard numbers, but
once in a while samba works and binds to both ipv4 and ipv6 at booting, but
mostly it only binds to ipv6. When it only binds to ipv6 and I do a "service
restart samba" samba wil come back after the restart bounded to both ipv4/ipv6.

However, I noticed some interesting thing in the log file /var/log/boot

--- boot--not_working.sorted	2010-05-15 21:54:39.000000000 +0200
+++ boot--working.sorted	2010-05-15 21:54:49.000000000 +0200
   Loading cpufreq kernel modules...done (acpi-cpufreq).
   Starting enhanced syslogd: rsyslogd.
   Starting Samba daemons: nmbd smbd.
- Starting web server: apache2apache2: apr_sockaddr_info_get() failed for mu-tm.student.utwente.nl
- apache2: Could not reliably determine the server's fully qualified domain name, using 127.0.0.1 for ServerName
- .
+ Starting web server: apache2.
   Turning on process accounting, file set to '/var/log/account/pacct'.
   Done..
   Checking battery state...done.

I did a "netstat -lnp |egrep "137|138|139|445" (with s/my real ip/myip/ to protect the innocent)
for
a)working samba
b)not working samba
c)not working samba followed by "service restart samba"


a)
tcp        0      0 0.0.0.0:445             0.0.0.0:*               LISTEN      2180/smbd 
tcp        0      0 0.0.0.0:139             0.0.0.0:*               LISTEN      2180/smbd 
tcp6       0      0 :::445                  :::*                    LISTEN      2180/smbd 
tcp6       0      0 :::139                  :::*                    LISTEN      2180/smbd 
udp        0      0 myip:137                0.0.0.0:*                           2277/nmbd 
udp        0      0 0.0.0.0:137             0.0.0.0:*                           2277/nmbd 
udp        0      0 myip:138                0.0.0.0:*                           2277/nmbd 
udp        0      0 0.0.0.0:138             0.0.0.0:*                           2277/nmbd

b)
tcp6       0      0 :::445                  :::*                    LISTEN      2154/smbd 
tcp6       0      0 :::139                  :::*                    LISTEN      2154/smbd 
udp        0      0 myip:137                0.0.0.0:*                           2361/nmbd 
udp        0      0 0.0.0.0:137             0.0.0.0:*                           2361/nmbd 
udp        0      0 myip:138                0.0.0.0:*                           2361/nmbd 
udp        0      0 0.0.0.0:138             0.0.0.0:*                           2361/nmbd

c)
tcp        0      0 0.0.0.0:445             0.0.0.0:*               LISTEN      4549/smbd 
tcp        0      0 0.0.0.0:139             0.0.0.0:*               LISTEN      4549/smbd 
tcp6       0      0 :::445                  :::*                    LISTEN      4549/smbd 
tcp6       0      0 :::139                  :::*                    LISTEN      4549/smbd 
udp        0      0 myip:137                0.0.0.0:*                           4547/nmbd 
udp        0      0 0.0.0.0:137             0.0.0.0:*                           4547/nmbd 
udp        0      0 myip:138                0.0.0.0:*                           4547/nmbd 
udp        0      0 0.0.0.0:138             0.0.0.0:*                           4547/nmbd

System: mixed testing/unstable.
kernel   : linux-image-2.6.32-5-amd64
samba    : 2:3.4.8~dfsg-1
netbase  : netbase
ifupdown : ifupdown
sysv-rc  : sysv-rc (using dependency based booting)

Thus far the facts, now my guestimate.

Could it be that sometimes ipv6 is up and running and ipv4 not yet, however
the boot process continues and services are started without ipv4?
Then when I manually restart a service, ipv4 is already up and ready, and by
restarting the service, it picks up the ipv4 sockets?

I'm willing to test patches,
regards

Martijn







-----------------------------------------
Ceterum censeo Carthaginem esse delendam