Public bug reported: Having reverted to having avahi allow user service publishing, and reverting afpd.conf to its default state to allow it to do so, netatalk does now correctly advertise itself to a Mac.
However, on trying to connect to it, the mac reports a "problem" (typically effusive), and in the ubuntu server's syslog this happens: Jun 9 16:30:36 twilight afpd[3215]: =============================================================== Jun 9 16:30:36 twilight afpd[3215]: INTERNAL ERROR: Signal 11 in pid 3215 (2.2.2) Jun 9 16:30:36 twilight afpd[3215]: =============================================================== Jun 9 16:30:36 twilight afpd[3215]: BACKTRACE: 3 stack frames: Jun 9 16:30:36 twilight afpd[3215]: #0 /usr/sbin/afpd(netatalk_panic+0x1c) [0x7f6df4b72b0c] Jun 9 16:30:36 twilight afpd[3215]: #1 /usr/sbin/afpd(+0x53c0c) [0x7f6df4b72c0c] Jun 9 16:30:36 twilight afpd[3215]: #2 /lib/x86_64-linux-gnu/libc.so.6(+0x364c0) [0x7f6df34cf4c0] Jun 9 16:30:50 twilight afpd[3222]: =============================================================== Jun 9 16:30:50 twilight afpd[3222]: INTERNAL ERROR: Signal 11 in pid 3222 (2.2.2) Jun 9 16:30:50 twilight afpd[3222]: =============================================================== Jun 9 16:30:50 twilight afpd[3222]: BACKTRACE: 3 stack frames: Jun 9 16:30:50 twilight afpd[3222]: #0 /usr/sbin/afpd(netatalk_panic+0x1c) [0x7f6df4b72b0c] Jun 9 16:30:50 twilight afpd[3222]: #1 /usr/sbin/afpd(+0x53c0c) [0x7f6df4b72c0c] Jun 9 16:30:50 twilight afpd[3222]: #2 /lib/x86_64-linux-gnu/libc.so.6(+0x364c0) [0x7f6df34cf4c0] Jun 9 16:30:53 twilight afpd[3226]: =============================================================== Jun 9 16:30:53 twilight afpd[3226]: INTERNAL ERROR: Signal 11 in pid 3226 (2.2.2) Jun 9 16:30:53 twilight afpd[3226]: =============================================================== Jun 9 16:30:53 twilight afpd[3226]: BACKTRACE: 3 stack frames: Jun 9 16:30:53 twilight afpd[3226]: #0 /usr/sbin/afpd(netatalk_panic+0x1c) [0x7f6df4b72b0c] Jun 9 16:30:53 twilight afpd[3226]: #1 /usr/sbin/afpd(+0x53c0c) [0x7f6df4b72c0c] Jun 9 16:30:53 twilight afpd[3226]: #2 /lib/x86_64-linux-gnu/libc.so.6(+0x364c0) [0x7f6df34cf4c0] The first one was on initially clicking on the host in the Mac's finder. The second one was after clicking on "Connect As" and submitting fresh login credentials, known to be correct. The third one was upon clicking OK in the error dialog that appears on the Mac. On the mac, at the same time, this is logged in kernel.log: 6/9/12 4:30:36.000 PM kernel: ASP_TCP CheckReqQueueSize: increasing req queue from 32 to 128 entries. so 0xffffff802041e8e0 6/9/12 4:30:36.000 PM kernel: ASP_TCP CancelOneRequest: cancelling slot 2 error 89 reqID 4 flags 0x9 afpCmd 0x13 so 0xffffff802041e8e0 6/9/12 4:30:51.000 PM kernel: ASP_TCP CheckReqQueueSize: increasing req queue from 32 to 128 entries. so 0xffffff8024d9ca38 6/9/12 4:30:51.000 PM kernel: ASP_TCP CancelOneRequest: cancelling slot 2 error 89 reqID 4 flags 0x9 afpCmd 0x13 so 0xffffff8024d9ca38 6/9/12 4:30:54.000 PM kernel: ASP_TCP CheckReqQueueSize: increasing req queue from 32 to 128 entries. so 0xffffff80186e8108 6/9/12 4:30:54.000 PM kernel: ASP_TCP CancelOneRequest: cancelling slot 2 error 89 reqID 4 flags 0x9 afpCmd 0x13 so 0xffffff80186e8108 Seems to be one pair for each internal error reported on the linux side. NB I don't think this is related to the earlier-reported-by-me bug #1010915 as I think that was avahi/zeroconf related and I believe I resolved that by reverting to default configuration. This involves the actual attempt to open an afp connection from a mac; zeroconf discovery seems to be working OK. Also, unlike the other bug, apport wasn't automatically triggered on the linux side; probably because the netatalk/afpd master process isn't being killed now, possibly only sub- processes. a changed afpd.conf is included below, but in fact every line is commented out, as per the original, so this should be a default configuration. (History is: I initially allowed the default version to be written during upgrade to quantal; then tried adding -nozeroconf and uncommenting that option line, then commented-out that option line again.) The "changed" AppleVolumes.default file is actually unchanged from the previous working Precise configuration. ProblemType: Bug DistroRelease: Ubuntu 12.10 Package: netatalk 2.2.2-1 ProcVersionSignature: Ubuntu 3.4.0-5.11-generic 3.4.0 Uname: Linux 3.4.0-5-generic x86_64 NonfreeKernelModules: nvidia ApportVersion: 2.1.1-0ubuntu2 Architecture: amd64 Date: Sat Jun 9 16:34:43 2012 InstallationMedia: Ubuntu 11.10 "Oneiric Ocelot" - Release amd64 (20111012) ProcEnviron: LANGUAGE=en_GB:en TERM=xterm PATH=(custom, user) LANG=en_GB.UTF-8 SHELL=/bin/bash SourcePackage: netatalk UpgradeStatus: Upgraded to quantal on 2012-06-09 (0 days ago) mtime.conffile..etc.netatalk.AppleVolumes.default: 2011-12-07T21:15:59.378054 mtime.conffile..etc.netatalk.afpd.conf: 2012-06-09T16:23:19.630006 ** Affects: netatalk (Ubuntu) Importance: Undecided Status: New ** Tags: amd64 apport-bug quantal running-unity -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1010921 Title: netatalk errors on trying to connect from real mac To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/netatalk/+bug/1010921/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs