OK. The patch I just sent might help too ... Best regards, Kern
On 03/31/2014 05:16 PM, Dan Langille wrote: > On 2014-03-31 10:41 AM, Kern Sibbald wrote: >> Hello Dan, >> >> Sorry I didn't read to the bottom of your email. Well, then the problem >> is very likely what you originally said -- a configuration error. >> Apparently, Bacula has detected that IPv6 is enabled, but it isn't >> really on your machine. >> Please look in <bacula>/src/config.h after having done a ./configure. >> >> If you have: >> >> #define HAVE_IPV6 1 > > We do have that. And just to show what we do have: > > ifconfig > em0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500 > > options=4219b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,TSO4,WOL_MAGIC,VLAN_HWTSO> > > ether 00:25:90:a4:cd:d4 > inet 10.55.0.19 netmask 0xffffffff broadcast 10.55.0.19 > media: Ethernet autoselect (1000baseT <full-duplex>) > status: active > em1: flags=8c02<BROADCAST,OACTIVE,SIMPLEX,MULTICAST> metric 0 mtu 1500 > > options=4219b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,TSO4,WOL_MAGIC,VLAN_HWTSO> > > ether 00:25:90:a4:cd:d5 > media: Ethernet autoselect > status: no carrier > lo0: flags=8049<UP,LOOPBACK,RUNNING,MULTICAST> metric 0 mtu 16384 > options=600003<RXCSUM,TXCSUM,RXCSUM_IPV6,TXCSUM_IPV6> > tun0: flags=8010<POINTOPOINT,MULTICAST> metric 0 mtu 1500 > options=80000<LINKSTATE> > > >> and there is no IPv6 on your machine, there is likely to be a problem. >> The quickie patch to see if that is the problem is to comment out that >> statement, build again and re-run the test. > > To do that test, I altered the configure script: > > $as_echo "#define HAVE_IPV6 0" >>confdefs.h > > Then restarted my regression test in the usual way. After > configuration completed, and while the test process was still running, > I checked from another shell: > > [dan@pg93:~/src/Bacula7RegressionTesting-TRUNK/regress/build/src] $ > grep IPV6 config.h > #define HAVE_IPV6 0 > >> What I don't understand is why TLS turns on IPv6. Perhaps OpenSSL uses >> it by default. If that is the case, then it is possible that turning >> off IPv6 in Bacula (as described above) will not resolve the problem, >> and the problem will then be in the OpenSSL libraries. > > If those tests run clear, I can try turning off TLS in the main > configuration, and see if that alters the value of IPV6 > >> >> Best regards, >> Kern >> >> >> On 03/31/2014 03:01 PM, Dan Langille wrote: >>> On 2014-03-31 07:36 AM, Kern Sibbald wrote: >>>> -----BEGIN PGP SIGNED MESSAGE----- >>>> Hash: SHA1 >>>> >>>> Hello Dan, >>>> >>>> I have to admit that I haven't run a Dart test for a very long time. >>>> The plugin-test should be removed from it, which I have done. I've >>>> pushed the patch to the git repo. >>>> >>>> Concerning the tls-test: that is a problem. Can you tell me what >>>> TCP/IP protocol 28 is on FreeBSD? >>> Nothing, that I see: >>> >>> $ grep 28 /etc/services | head >>> gss-xlicen 128/tcp #GSS X License Verification >>> gss-xlicen 128/udp #GSS X License Verification >>> http-mgmt 280/tcp >>> http-mgmt 280/udp >>> personal-link 281/tcp >>> personal-link 281/udp >>> cableport-ax 282/tcp #cable port a/x >>> cableport-ax 282/udp #cable port a/x >>> rescap 283/tcp >>> rescap 283/udp >>> >>>> This test runs perfectly here, and >>>> we have made some significant networking changes. This protocol should >>>> be the sa_family member and should normally be AF_INET for IPv4. >>>> However, now Bacula handles IPv6 much better than previous versions, >>>> and apparently there is a problem or difference with the FreeBSD IP >>>> definitions. For IPv6 this value should be AF_INET6. Those are the >>>> only two values that Bacula understands. >>> FYI, there is no IPv6 on the regression testing machine in question. >>> >>>> Note for Linux: AF_INET == 2 AF_INET6 = 10, so 28 is something >>>> different and doesn't even exist on Linux. >>> Guess what, you're right. See: >>> >>> https://www.freebsd.org/doc/en/books/developers-handbook/sockets-essential-functions.html >>> >>> >>> #define AF_INET6 28 >>> >>> >>>> Best regards, >>>> Kern >>>> >>>> On 03/31/2014 04:04 AM, Dan Langille wrote: >>>> > On Mar 30, 2014, at 4:44 PM, Dan Langille <d...@langille.org> wrote: >>>> > >>>> >> I tried some regression tests for Bacula 7 tonight. I found what I >>>> think may be a configure issue. >>>> > >>>> > The test finished, two errors: >>>> > >>>> > http://regress.bacula.org/buildSummary.php?buildid=24125 [1] >>>> > >>>> > >>>> > >>>> > >>>> ------------------------------------------------------------------------------ >>>> >>>> > >>>> > >>>> > _______________________________________________ >>>> > Bacula-devel mailing list >>>> > Bacula-devel@lists.sourceforge.net >>>> > https://lists.sourceforge.net/lists/listinfo/bacula-devel [2] >>>> >>>> -----BEGIN PGP SIGNATURE----- >>>> Version: GnuPG v1.4.11 (GNU/Linux) >>>> Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ [3] >>>> >>>> iEYEARECAAYFAlM5U0gACgkQNgfoSvWqwEi0CwCg3o5/4r1t4c4t+GhBroPBGUzN >>>> 9DEAoIdsqFwiAqHGJOyvzSsEc/5uENOh >>>> =R++t >>>> -----END PGP SIGNATURE----- >>>> >>>> >>>> >>>> Links: >>>> ------ >>>> [1] http://regress.bacula.org/buildSummary.php?buildid=24125 >>>> [2] https://lists.sourceforge.net/lists/listinfo/bacula-devel >>>> [3] http://www.enigmail.net/ > ------------------------------------------------------------------------------ _______________________________________________ Bacula-devel mailing list Bacula-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-devel
