stays in registering state after running

[seb128@…]

Using 0.6.22 on ubuntu hardy, avahi stays in registering state, using gdb shows it goes to running and then registering

(gdb) bt
#0  server_callback (s=0x8068a68, state=AVAHI_SERVER_RUNNING, userdata=0x8063d00) at main.c:342
#1  0xb7f8dcf0 in server_set_state (s=0x8068a68, state=AVAHI_SERVER_RUNNING) at server.c:1078
#2  0xb7f8e45a in avahi_host_rr_entry_group_callback (s=0x8068a68, g=0x804db40, state=134626560, userdata=0x0)
    at server.c:1118
#3  0xb7f90e87 in avahi_s_entry_group_change_state (g=0x807bb58, state=AVAHI_ENTRY_GROUP_ESTABLISHED) at entry.c:1039
#4  0xb7f9c6ff in avahi_s_entry_group_check_probed (g=0x807bb58, immediately=0) at announce.c:83
#5  0xb7f91075 in entry_group_commit_real (g=0x807bb58) at entry.c:1103
#6  0xb7f911ee in avahi_s_entry_group_commit (g=0x807bb58) at entry.c:1141
#7  0xb7f8b0b9 in avahi_interface_address_update_rrs (a=0x807bb20, remove_rrs=0) at iface.c:85
#8  0xb7f8b331 in avahi_interface_update_rrs (i=0x8063d00, remove_rrs=0) at iface.c:112
#9  0xb7f8b5ae in avahi_hw_interface_update_rrs (hw=0x807b000, remove_rrs=0) at iface.c:123
#10 0xb7f8baa1 in avahi_interface_monitor_update_rrs (m=0x8063d00, remove_rrs=0) at iface.c:168
#11 0xb7f8dcd3 in server_set_state (s=0x8068a68, state=134536000) at server.c:1075
#12 0xb7f8de08 in register_stuff (s=0x8068a68) at server.c:1209
#13 0xb7f90839 in avahi_server_new (poll_api=0x80663c8, sc=0x8063d00, callback=0x804db40 <server_callback>, 
    userdata=0x8063d00, error=0xbfed34e8) at server.c:1438
#14 0x0804e414 in run_server (c=0x8063d00) at main.c:1064
#15 0x0804f130 in main (argc=1, argv=Cannot access memory at address 0x4
) at main.c:1490
#16 0xb7de2450 in __libc_start_main () from /lib/tls/i686/cmov/libc.so.6
#17 0x0804cf71 in _start ()
(gdb) c
Continuing.
Server startup complete. Host name is seb128-desktop.local. Local service cookie is 3235680022.

Breakpoint 1, server_callback (s=0x8068a68, state=AVAHI_SERVER_REGISTERING, userdata=0x8063d00) at main.c:342
342	in main.c
(gdb) bt
#0  server_callback (s=0x8068a68, state=AVAHI_SERVER_REGISTERING, userdata=0x8063d00) at main.c:342
#1  0xb7f8dcf0 in server_set_state (s=0x8068a68, state=AVAHI_SERVER_REGISTERING) at server.c:1078
#2  0xb7f8de08 in register_stuff (s=0x8068a68) at server.c:1209
#3  0xb7f90839 in avahi_server_new (poll_api=0x80663c8, sc=0x8063d00, callback=0x804db40 <server_callback>, 
    userdata=0x8063d00, error=0xbfed34e8) at server.c:1438
#4  0x0804e414 in run_server (c=0x8063d00) at main.c:1064
#5  0x0804f130 in main (argc=1, argv=Cannot access memory at address 0x4
) at main.c:1490
#6  0xb7de2450 in __libc_start_main () from /lib/tls/i686/cmov/libc.so.6
#7  0x0804cf71 in _start ()

[trepidacious]

I also see this bug running under uClinux on a Blackfin processor. It's intermittent, and seems to vary considerably in probability of happening - when I first tested Avahi I saw no issues, but recently the daemon very often hangs in registering state as listed by "ps ax". It is also impossible to restart. I can provide more details if required. Would be great to see this resolved or at least responded to, since at the moment it makes Avahi too unreliable to use on an embedded device.

The same issue seems to crop up on Ubuntu as well - this looks similar:  https://bugs.launchpad.net/ubuntu/+source/avahi/+bug/116984

[fszczerba]

I see this on CentOS 5.4 with avahi-0.6.16-6.el5 and avahi-0.6.16-10.el5_6.

I can recreate it by running

service avahi-daemon restart ; service network restart

I've also seen it happen when a network interface happened to go down just as avahi was restarted.

When this happens, the log messages

avahi-daemon[10047]: Server startup complete. Host name is foo.local. Local service cookie is 2973069976.
avahi-daemon[10047]: Service "SFTP File Transfer on foo" (/services/sftp-ssh.service) successfully established.

which usually come out at the end of startup are not seen.