Re: [avahi] Is systemd planning to fully replace avahi?
On Di, 11.10.22 20:26, Petr Menšík (pemen...@redhat.com) wrote: > Hi, > > systemd has PR implementing [1] deeper mDNS functionality. I think that > means they intend to fully replace avahi by systemd-resolved. I don't like > that, but because very low activity in Avahi project, I understand that. > > Do we want to keep avahi primary choice for mDNS? What are you opinions on > replacing avahi with systemd-resolved? Yeah, this is definitely on my list. resolved already covers a lot of ground of mDNS/DNS-SD but there's quite some stuff missing still, most importantly API compat. Patches welcome! Lennart -- Lennart Poettering, Berlin
Re: [avahi] running 2 instances of avahi
On Wed, 05.09.12 13:38, Troyanker, Vlad (v...@lab126.com) wrote: I have a need to run avahi alongside running avahi-daemon. I have seen avahi printning a warning on startup when detecting non-exclusive use of its public port. However, it appears to work just fine (at least on ubuntu 11). My question is are there any limitations or restrictions in running 2 or more instances of avahi? Yes. It doesn't work for many messages, since unicast responses are generally only delivered to one listening socket, not all of them. Just don't do this. It's not supported. I added the message precisely to tell people about this problem. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Is it possible to get an A record from an Avahi server via IPv6?
On Tue, 31.07.12 19:45, jeremyf (jfrie...@gmail.com) wrote: Hi all, I've got a Linux server running avahi, and a client that manually sends out a simple, hand-rolled mDNS query packet over IPv6 (to ff02::fb) and parses the UDP packets it gets back to discover who is out there. This all works fine for my purposes, since most of the server software I use speaks IPv6. My problem is that there is a particular piece of server software running on the server that is only listening for incoming IPv4 connections -- if I try to connect to that server's port via IPv6, I get connection refused. My first response to this problem would be to fix the server to also listen on IPv6, but unfortunately in this case I'm not allowed to do that. My next impulse was to check the mDNS reply packets my client receives, to see if they contain an 'A' record that would tell me the server's IPv4 address. Then my client could make an IPv4 connection to that address. However, the mDNS reply packets I receive don't contain any 'A' records, only an '' record. So my question is, is there any special flag or something I can put in my mDNS query packet that would convince the Avahi server to include an 'A' record in its response? Or am I out of luck, and the only solution is to send mDNS query packets via both IPv4 and IPv6? AFAIR we sent that anyway in the packet in the auxiliary section. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] [PATCH 1/2] build-sys: fix parallel install in avahi-utils
On Sun, 17.06.12 07:44, Diego Elio Pettenò (flamee...@flameeyes.eu) wrote: The moment install-exec-local is called, we might still not have created ${DESTDIR}/${bindir} so we should make sure to create it first, and then try to chdir into it. Applied the first patch. Didn't apply the second patch because I already merged one doing the same. Thanks! Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] [PATCH 1/2] build-sys: fix parallel install in avahi-utils
On Tue, 19.06.12 16:37, Diego Elio Pettenò (flamee...@flameeyes.eu) wrote: Il 19/06/2012 15:48, Lennart Poettering ha scritto: Applied the first patch. Didn't apply the second patch because I already merged one doing the same. Make sure that the first one is changed to use MKDIR_P then, otherwise it'll fail. APplied another patch now, that fixes that. Thanks, Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Memory consumption
On Wed, 11.04.12 17:41, Frank Lahm (frankl...@googlemail.com) wrote: So it seems it's pthread_create(). Wonder why it allocates such a huge chunk of memory. I guess the only way avoiding this is completely redesigning our application such that the Avahi stuff is not done in the process that later forks and runs user AFP sessions. :( Would it make sense to extend the Avahi API so that a parameter could be passed which is then used to call pthread_attr_setstacksize() for the thread? As mentioned, this will only minimize address space, not memory usage. Allocating a biggre stack comes at virtually no cost -- as long as it is never accessed. Unless you turned off lazy allocation in Linux chasing this allocation is a waste of time. You are chasing a ghost. Have you turned off lazy allocation on Linux? (Note that the event loop logic in avahi is fully abstracted, you can implement your own logic, and run the stuff in your own thread with your own parameters. AvahiThreadedPoll is just one implementation, but nothing you actually have to use. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Memory consumption
On Wed, 11.04.12 16:50, Frank Lahm (frankl...@googlemail.com) wrote: I don't see where the Avahi client libraries cuould do such a big allocation. My only guess is that this is actually the D-Bus library that does this (it maintains a message cache). To figure this out it might be worth to plot the memory usage with a tool like valgrind's massif tool? I had checked with Valgrind memcheck before to no avail. I've looked at massif as to your recommendation and that is indeed quite revealing (this requires the --pages-as-heap=yes option as the large allocation is apparently not done via malloc et al): So it seems it's pthread_create(). Wonder why it allocates such a huge chunk of memory. I guess the only way avoiding this is completely redesigning our application such that the Avahi stuff is not done in the process that later forks and runs user AFP sessions. :( This might simply be the stack for the thread, which you can influence with ulimit -s in its size. However, note that having a large stack doesn't actually mean that this uses a lot of memory, it just reserves address space, which is only backed when needed. Hence, most likely this is really a non-issue. (i.e. don't confuse reservation of memory with reservation of address space!) Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Memory consumption
On Tue, 10.04.12 15:23, Frank Lahm (frankl...@googlemail.com) wrote: Hi all, I'm one of the Netatalk (OpenSource AFP fileserver) devs. We use Avahi via avahi_threaded_poll and friends in order to register the AFP server with mdns. We call the Avahi setup stuff once in our main process `afpd` here [1]. Handling user sessions is done by forking childs. As it seems, when calling the Avahi setup functions, a big chunk of memory is taken: Obviously, we didn't bother calling the Avahi free ressources() functions in the forked afpd session childs as the damage done by sbrk() can't be undone by free(). The main concern I have with this memory consumption is that is handed down to every afpd process. My questions are: - is the memory consumption of 10 MB as expected ? - do you see any way of preventing the inheritance of the memory consumption without running the Avahi registration in an dedicated process ? I don't see where the Avahi client libraries cuould do such a big allocation. My only guess is that this is actually the D-Bus library that does this (it maintains a message cache). To figure this out it might be worth to plot the memory usage with a tool like valgrind's massif tool? Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
[avahi] [HEADS-UP] Dropping avahi-qt3, avahi-ui for gtk2, avahi-howl
Heya, I am planning to drop Qt3 and Gtk2 support from avahi soon, leaving only Qt4 and Gtk3 in. Anybody opposed to that? You need a very strong case if you want to convince me not to do this. Similar I want to drop the HOWl compat kludge. Would like to drop the Bonjour compat kludge too, but I figure too many folks still use that. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
[avahi] [ANNOUNCE] Avahi 0.6.31
Heya, http://avahi.org/download/avahi-0.6.31.tar.gz This mostly fixes one thing only: we can build Avahi again with current autoconf. Lennart Poettering (5): systemd: syslog.target is not longer useful build-sys: compatibility with newer automake build-sys: prepare release update NEWS build-sys: actually, we need to place the service type database in lib64, not unconditionally Ludwig Nussel (1): database: add Mumble to service type database Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi daemon Withdrawing address record
On Tue, 05.07.11 17:41, Brandon Stafford (bran...@rascalmicro.com) wrote: Hi, I'm working on debugging an embedded Linux board that uses avahi 0.6.25 for mDNS. When DHCP succeeds, local name resolution works as expected. When DHCP fails, as it does when my board is connected directly to my laptop, it appears that avahi successfully claims a link-local address, but then immediately throws it away. Note that avahi-autoipd implements IPv4LL and avahi-daemon does not know much about it, it just watches netlink. Here's what I see in /var/log/messages: Jul 1 04:57:06 rascal14 daemon.info avahi-daemon[602]: Joining mDNS multicast group on interface eth0.IPv4 with address 169.254.124.120. Jul 1 04:57:06 rascal14 daemon.info avahi-daemon[602]: New relevant interface eth0.IPv4 for mDNS. Jul 1 04:57:06 rascal14 daemon.info avahi-daemon[602]: Registering new address record for 169.254.124.120 on eth0.IPv4. Jul 1 04:57:06 rascal14 daemon.info avahi-daemon[602]: Withdrawing address record for 169.254.124.120 on eth0. Jul 1 04:57:06 rascal14 daemon.info avahi-daemon[602]: Leaving mDNS multicast group on interface eth0.IPv4 with address 169.254.124.120. Jul 1 04:57:06 rascal14 daemon.info avahi-daemon[602]: Interface eth0.IPv4 no longer relevant for mDNS. These log messages are getting created by avahi-core/iface.c, but I'm not sure why avahi daemon would claim and then release an address. From what I can tell, the actual link-local address negotiation is handled by /usr/sbin/zeroconf, which is actually zeroconf 0.9 from http://www.progsoc.org/~wildfire/zeroconf/, but it's not clear to me how or if that's communicating with avahi-daemon. Something happened so that Avahi didn't consider the interface relevant anymore. This could be because somebody removed the IPv4LL address again. To find out it would be useful if you pasted the output of ip addr and ip link after this happened. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi daemon Withdrawing address record
On Tue, 05.07.11 18:49, Brandon Stafford (bran...@rascalmicro.com) wrote: Something happened so that Avahi didn't consider the interface relevant anymore. This could be because somebody removed the IPv4LL address again. To find out it would be useful if you pasted the output of ip addr and ip link after this happened. [root@rascal14$:~]: ip addr 1: lo: LOOPBACK,UP,LOWER_UP mtu 16436 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo 2: eth0: BROADCAST,MULTICAST,UP,LOWER_UP mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 02:71:82:06:00:14 brd ff:ff:ff:ff:ff:ff [root@rascal14$:~]: ip link 1: lo: LOOPBACK,UP,LOWER_UP mtu 16436 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 2: eth0: BROADCAST,MULTICAST,UP,LOWER_UP mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 02:71:82:06:00:14 brd ff:ff:ff:ff:ff:ff This is after executing /etc/init.d/networking restart, which induces the same series of log messages I pasted earlier. This confirms that eth0 is missing an inet address, but it doesn't give me (in my state of ignorance) any additional insight. By somebody removed the IPv4LL address again, I assume you mean a process other than avahi-daemon. Might udhcpc be a candidate? It's also running. I don't see any other processes running that are relevant to networking, at least as far as I can tell. Yes, could be udhcpc. Could be avahi-autoipd too, for example, when you terminate it, ot wehn it encounters an IP conflict. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] compile avahi on CentOS 5
On Wed, 29.06.11 17:39, valerie.bau...@bull.net (valerie.bau...@bull.net) wrote: Hello I'm working on a centOs 5 and the only rpm I found for avahi is quite old (0.6.16). With this version I can't find the option --parsable for avahi-browse, I suppose it comes with more recent versions So I would like to compile it by myself, but I'm not very familiar with linux compile ! What should I run after downloading the source ? I tried autogen.sh, but it gives only the following result and then seems to bug : + '[' -f .git/hooks/pre-commit.sample -a '!' -f .git/hooks/pre-commit ']' + '[' x = xam ']' + rm -rf autom4te.cache + rm -f config.cache + rm -f Makefile.am~ configure.ac~ ++ which gettextize + sh -s -- --copy --force which: no gettextize in (/usr/kerberos/sbin:/usr/kerberos/bin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/root/bin) + sed 's/read dummy/\#/' Can somebody help ? Install the gettext RPM. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Wide-Area Write Support and Long Lived Queries
On Wed, 15.06.11 09:40, Andrew Tunnell-Jones (and...@tj.id.au) wrote: On 05/06/2011, at 11:51 AM, Andrew Tunnell-Jones wrote: Hi folks, From what I can glean from avahi.org it looks like writable wide-area support and LLQ were looked at some years back but the work wasn't finished. The page for milestone 0.7[1] suggests these features are coming in that release, however as 0.6 has been around quite a few years and the page also references changes warranted by an RFC updated in 2005, I'm not sure how relevant that page is. Could someone let me know what the status of these features is? Thanks, Andrew 1. http://avahi.org/milestone/Avahi%200.7 Has anyone got an answer? Is there anyone else interested in enhancements to wide-area support? LLQ support is not available right now. Is server support for LLQ available in any unpached DNS server so far? (this was the reason I never bothered to implement it.) Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi does not show upnp resources
On Sat, 28.05.11 06:44, John Frankish (j-frank...@slb.com) wrote: Although avahi-browse shows daap resources, it does not seem to show upnp media server resources, such as rygel, streaming music files. Is this something avahi is meant to recognize? Avahi implements mDNS, DNS-SD, unicast DNS and IPV4LL. It does not implement UPnP. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Looking for help/advice
On Mon, 09.05.11 15:26, Robert Schweikert (rschweik...@novell.com) wrote: Hi, Trying again to find some help. I would like to automate certain tasks for machine configuration in a cluster setup. One example of this automatic setup would be setting up and NFS mount. Basically the head node of the cluster gets installed and manually configured. Then I'd like to run some kind of service on the head node and when a cluster node gets installed, upon firstboot some client code runs to find the service running on the head node. Then gets the IP address from the head node and the path for the NFS share it exports. Avahi is a tool for implementing distributed service discovery in peer to peer networks. You setup seems to be inherently client/server, hence my suggestion would be to configure the NFS server via DHCP on the clients. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] [PATCH] add Mumble to service type database
On Fri, 06.05.11 10:47, Ludwig Nussel (ludwig.nus...@suse.de) wrote: http://mumble.sourceforge.net/ --- Thanks! Applied. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Questions regarding service types
On Fri, 08.04.11 15:19, Nathanael D. Noblet (nathan...@gnat.ca) wrote: On 04/08/2011 07:24 AM, Lennart Poettering wrote: On Thu, 07.04.11 09:28, Nathanael D. Noblet (nathan...@gnat.ca) wrote: The database is used only and exclusively by avahi-browse and avahi-discover. If you need your own private additions then feel free to extend the database locally, but keep in mind that you need to rebuild the binary version of it with that Python script. Okay, so then does it matter? The database is useful only to make things pretty. It has no technical purpose beyond that. I won't be using tahose apps, I was just using them as a way to test what was going on. Do other apps that look for published services use their own inhouse db? Or do they just search for particular service types and ignore the rest (in which case I don't need to worry about them showing up elsewhere)... ? If you make up a new service type and believe it wil be used widely, then yeah, we can register it in the upstream db. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
[avahi] [ANNOUNCE] Avahi 0.6.30
Heya, http://avahi.org/download/avahi-0.6.30.tar.gz Changes: Lennart Poettering (9): systemd: drop ExecReload= line since avahi-dnsconfd -r is not actually a config reload general: various smaller fixes systemd: fix sd_is_socket() invocation iface: look for both IFA_ADDRESS and IFA_LOCAL Merge remote-tracking branch 'origin/master-tx' i18n: run make update-po iface: use interface values only when valid socket: minor beautification doc,build-sys: prepare new release Robert Ancell (1): avahi-discover: Fix type being inserted into tree model (Bug #332) Ville-Pekka Vainio (1): l10n: Updated Finnish (fi) translation to 99% el_libre como el chaval (1): l10n: Updated Catalan (ca) translation to 100% sir_ade (1): l10n: Updated Malay (ms) translation to 24% Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] problems with two interface setup
On Wed, 09.03.11 08:36, Alexey Fisher (bug-tr...@fisher-privat.net) wrote: after i switch eth0 off (pull off the cable), avahi stops announcing _presence_tcp, or at least do not re announce it on other interface/ip, even if empaty is listening on wlan0 too. I can force and send directly salut/xmmp message to emapty for testing, so the problem should be on avahi part. Your thoughts? If wait for a while, does the problem fix itself? This question sound sarcastic, but it is probably not what you mean :D Nope, it was actually meant the way I said it, i.e. that the TTL handling should fix it. No, the problem do not fix itself. Even if i wait one day. I presume the wlan and the cable ethernet are the same network actually? Yes. Can you check if you have an sylsog output related to collisions? Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi logs Host name conflict, retrying ..
On Thu, 10.03.11 02:08, Harikrishna Donti (harikrishna.do...@lntemsys.com) wrote: Hi All, I'm running avahi deamon on my boards and on some boards, it keeps logging Host name conflict, retrying with MYNGP0008F80201CE-XX. I'm sure no one else has that host name in my LAN. And with this the services on these boards are not discoverable. Can any one please help understanding and solving this? Moste likely due to some reason something on your network is send back all messages avahi send out and Avahi fails to recognize that it is looking at its own messages. Wireshark might help to track this down. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] New version of the LFS initscript fix patch
On Sun, 16.05.10 14:08, William Immendorf (will.immend...@gmail.com) wrote: I've attached a new vesrion of the Avahi LFS initscript patch, that's diffrent that the one that I've submitted a long time ago, which includes fixes so that it's more useful. If you would merge this into the next release, that would be gladly apprecated. +install: install-am + ln -sf ../init.d/avahi ${DESTDIR}/etc/rc.d/rc0.d/K28avahi + ln -sf ../init.d/avahi ${DESTDIR}/etc/rc.d/rc1.d/K28avahi + ln -sf ../init.d/avahi ${DESTDIR}/etc/rc.d/rc2.d/K49avahi + ln -sf ../init.d/avahi ${DESTDIR}/etc/rc.d/rc3.d/S21avahi + ln -sf ../init.d/avahi ${DESTDIR}/etc/rc.d/rc4.d/S21avahi + ln -sf ../init.d/avahi ${DESTDIR}/etc/rc.d/rc5.d/S21avahi + ln -sf ../init.d/avahi ${DESTDIR}/etc/rc.d/rc6.d/K28avahi I am not convinced this is a good idea: it is not an upstream decision to enable a service by default. That is definitely something to leave to the distro or the admin. I am happy to merge a patch with this part removed. (Sorry for the late reply, currently going through all unanswered mails to prep a new release) Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi_entry_group_free misbehaving
On Tue, 23.11.10 11:36, David Thompson (dcth...@sandia.gov) wrote: Hi, It appears that avahi_entry_group_free is multiply defined: once in avahi-client/group.c and once in avahi-core/entry.c. The problem is that the signatures are different and, at least on Fedora 13 with Avahi 0.6.25-7.fc13, the wrong signature is called from within avahi_client_free if both avahi-core and avahi-client are linked to a binary. I was accidentally linking to avahi-core (and unaware that it would cause problems in any event) and ran into this issue. Perhaps it makes sense to add a warning to http://www.avahi.org/wiki/ArchitecturalOverview about linking to both libraries, or at least about getting the search order right? Thanks, added that now. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] reflector over ipv6
On Sun, 02.01.11 09:29, bombela (bomb...@gmail.com) wrote: I hope this version fit better the coding style. Thanks! Applied. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi on Linux uses incorrect address for P-t-P interface
On Sun, 16.01.11 10:51, Matt Ginzton (m...@cs.stanford.edu) wrote: Hi, Heya, I sent a less knowledgeable question about avahi-daemon and point-to-point links a few days ago, http://lists.freedesktop.org/archives/avahi/2011-January/001969.html. When I didn't get a response to this, I decided to build avahi from source and step through it and see how it builds its list of interfaces and their addresses. Thanks a lot. Applied. Sorry for the long delay! The reason why I had IFA_ADDRESS there is because the rtnetlink(7) man page is not helpful in this regarding, and since we don't really support ptp links anyway nobody ever noticed. --- iface-linux.c.orig2011-01-16 10:43:19.928941366 -0800 +++ iface-linux.c 2011-01-16 10:44:53.802731728 -0800 @@ -203,7 +203,7 @@ while (RTA_OK(a, l)) { switch(a-rta_type) { -case IFA_ADDRESS: +case IFA_LOCAL: /* Fill in address data */ if ((raddr.proto == AVAHI_PROTO_INET6 RTA_PAYLOAD(a) != 16) || ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] problems with two interface setup
On Sat, 05.03.11 15:08, Alexey Fisher (bug-tr...@fisher-privat.net) wrote: Hallo all, i try to track down one avahi/empathy problem which i have with ubuntu nutty setup on my netbook. Empathy stops to update salut status after switching lan to wlan. I don not know where is the problem, on empathy or avahi site. Usually it satrt as fallow: boot, eth0 up, avahi announce eth0, gdm login, epathy, _presence_tcp announce on eth0, wlan0 up (same network as eth0) after i switch eth0 off (pull off the cable), avahi stops announcing _presence_tcp, or at least do not re announce it on other interface/ip, even if empaty is listening on wlan0 too. I can force and send directly salut/xmmp message to emapty for testing, so the problem should be on avahi part. Your thoughts? If wait for a while, does the problem fix itself? I presume the wlan and the cable ethernet are the same network actually? Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Eth0 2 IP's - resolves to 169 series IP and not to DHCP IP
On Wed, 09.03.11 00:38, Harikrishna Donti (harikrishna.do...@lntemsys.com) wrote: Hi, I have avahi installed on an ARM based board. I have my eth0 configured to DHCP and eth0 also has an Alias IP 169.254.1.100. The idea behind having an alias IP was an installer can always connect an Ethernet cable between his laptop and board and talk to board. When I start avahi, it logs messages . avahi-daemon 0.6.22 starting up WARNING: No NSS support for mDNS detected, consider installing nss-mdns! Loading service file /etc/avahi/services/MY.service Joining mDNS multicast group on interface eth0.IPv4 with address 169.254.1.200. New relevant interface eth0.IPv4 for mDNS Network interface enumeration completed. Registering new address record for 169.254.1.200 on eth0.IPv4. Registering new address record for 10.113.25.45 on eth0.IPv4 Hmm, how come one line has a final fulls top on it, and the other does not? Server startup complete. Host name is MY0008F80201CE.local. Local service cookie is 2388707700 When I try to discover the board, the host name most of the times resolves to 169.254.1.200 and not to the dhcp IP ( 10.112.25.45). I would like to discover or resolve to DHCP IP or both IP's. How can I achieve this? Is there any configuration not to discover 169.* IP's? Uh, this is weird. Normally avahi should just pick a single address per interface, and it should be the most global one. Is this on Linux? Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] New version of the LFS initscript fix patch
On Tue, 08.03.11 16:50, William Immendorf (will.immend...@gmail.com) wrote: On Tue, Mar 8, 2011 at 3:25 PM, Lennart Poettering lenn...@poettering.net wrote: I am not convinced this is a good idea: it is not an upstream decision to enable a service by default. That is definitely something to leave to the distro or the admin. I am happy to merge a patch with this part removed. Well then, you can just remove that part. It just means that when I update the BLFS book, I'll add the symlink instructions manually. May I ask you to prep a clean git patch? Thansk you, Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
[avahi] [ANNOUNCE] Avahi 0.6.29
Heya, Mostly bugfixesand a fix for CVE-2011-1002 which fixes the borked fix for CVE-2010-2244: http://avahi.org/download/avahi-0.6.29.tar.gz François-Xavier Bourlet (1): reflector: cache entry link-local addresses filter John Bradshaw (2): man: spelling fixes man: more spelling fixes Lennart Poettering (7): systemd: update sd-daemon.[ch] daemon: we better stay away from SIGQUIT since it is supposed to cause a core dump avahi-ui: gtk3 fix chroot: suppress error messages if the chroot pipe is closed systemd: update sd-daemon.[ch] build-sys: update package version and soname update README Matt Ginzton (1): core: avahi on Linux uses incorrect address for P-t-P interface Pablo Neira Ayuso (1): netlink: fix wrong use of netlink flags for dump operations Paddy Smith (1): setproctitle: use defined() around __linux__ in #elif Stef Walter (1): gobject: ga_service_browser_new() should accept const gchar* Ville-Pekka Vainio (1): l10n: Updated Finnish (fi) translation to 99% Vincent Untz (4): howl: fix small issue init: update the suse initscripts socket: Still read corrupt packets from the sockets init: suse, avahi-dnsconfd status reports dead instead of unused Zbigniew Jędrzejewski-Szmek (1): Keep emacs configuration in one configuration file. el_libre como el chaval (1): l10n: Updated Catalan (ca) translation to 100% sir_ade (1): l10n: Updated Malay (ms) translation to 24% Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] LLMNR?
On Thu, 24.02.11 19:14, Sean Simmons (s...@low-five.com) wrote: Hi, Digging through the list archives I ran across this message http://lists.freedesktop.org/archives/avahi/2008-December/001534.html and this message http://lists.freedesktop.org/archives/avahi/2009-April/001623.html that seem to indicate that support for LLMNR in Avahi was imminent at the time. Looking through the code I don't see any evidence of this and there have been several releases since that last message. Is this still on the menu? If not, can someone point me to some patches? I have unfortunately never found the time to review those patches. I take the full blame for this. Sorry, Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi-daemon segfault
On Wed, 09.02.11 23:31, Jonas Aamodt Moræus (mor...@barracuda.homelinux.org) wrote: Hello, I've been running avahi-daemon for a long time on debian stable, but after resent upgrade the avahi-daemon started to segfault when a Mac OSX machine connects to one of my shares over netatalk (AFP), with this written to syslog: How do I start debugging this? Please generate a backtrace via gdb (most distros have a wiki page about that, for example fedora: http://fedoraproject.org/wiki/StackTraces) and submit it into you distribution's bug tracker. The downstream maintainers will then escalate this to upstream if necessary. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi + CUPS
On Sat, 25.12.10 18:37, Adam Tauno Williams (awill...@whitemice.org) wrote: On Thu, 2010-12-09 at 05:29 -0500, Brian Cardarella wrote: I'm sure this has been asked before It has. By me [two years ago?] and at least once in the last six months by someone else. First see if your CUPS is linked to Avahi ldd /usr/sbin/cupsd | grep avahi Note that older CUPS are probably using the Avahi compat stuff for Apple's libdnssd ad hence this call will most likely not do what you expect. IIRC you need to grep for libdns_sd. Only newer CUPS is able to link directly against Avahi. (Or actually, this might be a patch fedora ships, but which is not included upstream, dunno). Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi and IPv6
On Fri, 31.12.10 15:13, W. Michael Petullo (m...@flyn.org) wrote: I have read several documents on the Internet that indicate that Avahi supports IPv6. I am using Fedora 14 and am not getting the behavior I expect from Avahi on IPv6. I have written use-ipv6=yes to avahi-daemon.conf, but when I run avahi-browse, I do not see any IPv6 services listed (only IPv4). For an additional test, I turned off use-ipv4. In this case, avahi-browse listed no services. I would expect to see at least the _workstation._tcp service listed as available on IPv6. Note that there is no such thing as an IPV6 service. There are two things: a) You use IPV6 as a protocol to exchange mDNS packets with and/or b) The hostnames in mDNS contain records, i.e. IPv6 address records. A service will never refer to a specific address, it only refers to a hostname, which might or might not have an A and/or record assigned. The a) listed above you may control via use-ipv6=yes/no. With publish--on-ipv4=yes you can influence b). Finally, I enabled ipv6, disabled ipv4 and ran avahi-publish -s test _test._tcp 80. This command responded with Established under name 'test', but avahi-browse did not display the service. Not at all? This would be an indication that your machine does not deliver IPv6 packets, in wich case you might want to use tcpdump or wireshark to see if this works. I have tried turning off my host firewall and have disabled SELinux. Are you sure you disabled ip6tables? It is orthogonal of iptables, which is IPv4 only. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] reflector over ipv6
On Tue, 21.12.10 18:02, bombela (bomb...@gmail.com) wrote: This patch correct the issue by filtering caches entries before sent them back to the querier. All ipv6 link-local addresses are simply ignored. I'd be willing to merge this patch. However, before I do this, could you fix the patch to follow the coding style of the rest? No // coments, no variable definitions in the middle of a block and proper indenting, please? Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi + CUPS
On Thu, 09.12.10 05:29, Brian Cardarella (bcardare...@gmail.com) wrote: I'm sure this has been asked before so I apologize (I've spent the last few hours looking for an answer but cannot find one) Anyway, I have a printer being shared with Bonjour. My server has CUPS (recent build) and Avahi installed. I can very see the printer with avahi-browse: + tap0 IPv4 Lexmark International Lexmark E260d @ Brian Cardarella’s MacB Internet Printer local But CUPS cannot discover the printer. Here is my CUPS config for browsing shared printers: Have you checked that your CUPS was built with Avahi/Bonjour support enabled? Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Unicast Responses
On Sat, 13.11.10 17:17, Janaína Lemos (janaina.le...@gmail.com) wrote: Heya, I'm doing some tests with avahi (I'm planning to use it in my dissertation) and would like to know if Avahi implements the item 5.5 in the mDNS draft, to allow the use of questions requesting unicast responses. This feature would be very useful for me, because i need answers to be sent in unicast for security reasons. Yes, Avahi honours the unicast response bit. Just out of curiosity, what's your dissertation going to be about? Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Reason for entry groups
On Mon, 25.10.10 13:43, Christiano F. Haesbaert (haesba...@haesbaert.org) wrote: Hi Trent, Thanks for the clarification. So it would make sense only when you have a group of services that you'd like to be published under the same name. Yes. You'll get a single conflict notification for the entry group hence it would be a bad idea to add multiple services with different names to it, simply because you wouldn't know which one of it is actually conflicting. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Compilation issue
On Thu, 21.10.10 11:47, Gautham (gaut...@mindtree.com) wrote: Hello I'd like to run avahi just to register a static service on the device. The device is running montavista linux (2.6.18 kernel). I'm using 0.6.28 verson of avahi source. While configuring I'm disabling the following options. ./configure --disable-qt3 --disable-qt4 --disable-gtk --disable-gtk3 --disable-gdbm --disable-python --disable-pygtk --disable-python-dbus --disable-glib --disable-gobject --disable-nls --disable-dbus The .configure is successful then I run make. I get the following error while running make. sd-daemon.c: In function âsd_notifyâ: sd-daemon.c:327: warning: unused parameter âunset_environmentâ [-Wunused-parameter] sd-daemon.c:327: warning: unused parameter âstateâ [-Wunused-parameter] sd-daemon.c: In function âtouchâ: sd-daemon.c:445: error: âO_CLOEXECâ undeclared (first use in this function) sd-daemon.c:445: error: (Each undeclared identifier is reported only once sd-daemon.c:445: error: for each function it appears in.) make[2]: *** [avahi_daemon-sd-daemon.o] Error 1 make[2]: Leaving directory `/root/gautham/avahi-0.6.28/avahi-daemon' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/root/gautham/avahi-0.6.28' make: *** [all] Error 2 Any clues? This happens if you have a very old glibc. This is fixed in Avahi git. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Problems in compiling
On Tue, 05.10.10 07:30, Gautham (gaut...@mindtree.com) wrote: Actually I need a basic avahi daemon that registers a static service in /etc/services. Do I need to install Qt for running this as well. Am I missing anything? Thanks. Please, write proper emails to this list. No HTML please. And I already answered your question. http://lists.freedesktop.org/archives/avahi/2010-September/001934.html Like almost any other dep of Avahi you can disable the qt dep with a switch to configure, in this case --disable-qt4. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
[avahi] [ANNOUNCE] Avahi 0.6.28
Heya, http://avahi.org/download/avahi-0.6.28.tar.gz This is a bugfix release. * Updates regarding systemd integration * Properly avoid bus activation on non-systemd systems * Compatibility with newer gtk3 and gobject introspection * i18n updates * Minor other updates This release is backwards compatible with Avahi 0.6.x with x 28. Andika Triwidada (3): l10n: Initial Indonesian translation l10n: Updated Indonesian (id) translation to 97% l10n: Updates to Indonesian translation Bruno_Patri (1): l10n: Updated French (fr) translation to 99% Cheng-Chia Tseng (1): l10n: Updated Chinese (Taiwan) (zh_TW) translation to 66% Daniel Nylander (1): l10n: Updated Swedish (sv) translation to 100% Gunleif Joensen (1): l10n: Preliminary faroese translation Joe Hansen (3): l10n: new danish translation l10n: Updated Danish (da) translation to 100% l10n: Updated Danish (da) translation to 100% Lennart Poettering (10): gcc: minor enum issues gcc 4.5 discovered build-sys: require libdaemon 0.14 -- see #317 Merge remote branch 'origin/master-tx' i18n: update LINGUAS systemd: update sd-daemon.[ch] systemd: pull in avahi from dbus activation indirectly to make it possible to deactivate it client: use Ping() to test whether avahi is around bump version update NEWS systemd: update configure.ac snippet Luca Ferretti (1): l10n: Updated Italian (it) translation to 99% Matthias Clasen (1): gtk: avahi build fix Takayuki KUSANO (1): l10n: Updated Japanese (ja) translation to 100% Vincent Untz (1): gobject: Fix avahi build with gobject-introspection Yaron Shahrabani (1): l10n: Updated Hebrew translation by Yaron Shahrabani. Yuri Kozlov (1): l10n: Updated Russian (ru) translation to 100% kelemeng (2): l10n: Updated Hungarian (hu) translation to 98% l10n: Updated Hungarian (hu) translation to 100% mateju (3): l10n: Updated Slovenian translation. l10n: Updated Slovenian translation. l10n: Updated Slovenian translation milo (1): l10n: Updated Italian (it) translation to 100% rpress (1): l10n: Updated German (de) translation to 98% Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi Digest, Vol 66, Issue 6
On Wed, 29.09.10 09:09, Gautham (gaut...@mindtree.com) wrote: Thanks Lennart for the reply. I downloaded the latest source avahi-0.6.27.tar.gz unzipped ran the command ./configure --sysconfdir=/etc --localstatedir=/var --disable-gtk --disable-dbus --disable-python on my Linux Box ( See the details of the system below) [r...@a2md02576 avahi-0.6.27]# cat /proc/version Linux version 2.6.18-53.el5 (brewbuil...@hs20-bc2-3.build.redhat.com) (gcc version 4.1.2 20070626 (Red Hat 4.1.2-14)) #1 SMP Wed Oct 10 16:34:02 EDT 2007 However I get an error saying checking for QT4... configure: error: Package requirements ( QtCore = 4.0.0 ) were not met: No package 'QtCore' found Consider adjusting the PKG_CONFIG_PATH environment variable if you installed software in a non-standard prefix. Alternatively, you may set the environment variables QT4_CFLAGS and QT4_LIBS to avoid the need to call pkg-config. See the pkg-config man page for more details. Do I need to install Qt for running basic avahi daemon that registers a service in /etc/services like you mentioned previously? Am I missing anything? Thanks. No. Qt is not needed. There's a table on http://avahi.org/wiki/DownloadAvahi what the various deps give you. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Newbie Question
On Thu, 23.09.10 13:19, Gautham (gaut...@mindtree.com) wrote: I'm new to Avahi just going through the readme the source to understand what's going on. I'm planning to put avahi on one our devices running montavista Linux 2.6.18. I need only service registration (service publishing) so that my windows client can discover my device. Can I take only the avahi-daemon avahi-common x-compile to make it work? As in almost every other free software project use switches of the configure script to configure which options you want and which ones you don't. Note that the local browsing and registration client interface is exposed via D-Bus. That means if you want to write a program that publishes services via Avahi you need at least D-Bus enabled. However, if all you need is defining one static service you can rely on /etc/avahi/services and place a static service definition files there in which case you don't need D-Bus. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] How to add additional RRs in mdns response package?
On Thu, 02.09.10 13:14, 張皓茗 (diodr...@gmail.com) wrote: Hello I'm new to avahi and I'm learning it. I have some question about construct some network message between bonjour server and client. I would like to add some additional RRs in mdns response package, and I add these function avahi_record_new_full(), avahi_server_add() and avahi_record_unref() to make it while create service, in create_services(), does it a wrong way? my english is not well, with you can understand the mail and what I mean. As an Avahi client you may call avahi_entry_group_add_record() to publish a naked RR. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Binary data in TXT records
On Tue, 10.08.10 07:40, Jeff Koftinoff (je...@jdkoftinoff.com) wrote: I just read the following from http://www.zeroconf.org/Rendezvous/txtrecords.html which states that while it is a TXT record, the NAME is ASCII, but the VALUE is 8 bit binary and you SHOULD NOT convert binary values into hexadecimal ascii! What is the best way to encode binary values in a TXT record as defined by the xml service file in avahi? Does avahi handle embedded nulls in the value? The C API allows you to create binary TXT fields very easily. However, this is not supported in the XML definition for static services, simply because there is no nice way to include binary data in XML files. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi 0.6.27
On Wed, 04.08.10 16:28, Stanislav Gutev (stanislav.gu...@dfki.de) wrote: Hello, could someone please explain to me what is the reason for the following change in avahi-client/client.c: 562,564c563 if (!(dbus_bus_name_has_owner(client-bus, AVAHI_DBUS_NAME, error)) || dbus_error_is_set(error)) { --- if (!dbus_bus_start_service_by_name(client-bus, AVAHI_DBUS_NAME, 0, NULL, error)) { For some reason, when using the second method, the client doesnt find the running daemon, but if I try it with the previous one, it works properly. Thank you in advance. Well, we changed this in context of systemd: we want to make it possible to activate avahi only on demand, i.e. when at least one of the following three triggers are triggered: 1) a dbus client talks to the service 2) a network is configured 3) somebody connects to the simple protocol socket To make this possible we had to change a few things and instead of having the client libs check whether avahi is around we just activate it and see if this works. There are three possible fixes: 1) the dirty one, just install org.fd.Avahi.service into /usr/share/dbus-1/system-services/ even if you are not running systemd. 2) fix d-bus to not return an error when we try to manually activate a service that is running but not activatable 3) fix avahi-client to just call org.fd.DBus.Ping on the service instead of anything else, which should do what is requested. Hppy to take patches for #3. Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
[avahi] [ANNOUNCE] Avahi 0.6.27
Heya! Just rolled another release: http://avahi.org/download/avahi-0.6.27.tar.gz Héctor Daniel Cabrera (1): l10n: Updated Spanish (Castilian) (es) translation to 100% Lennart Poettering (17): systemd: add description strings to all systemd services/sockets systemd: simplify systemd unit files following the recent DefaultDependencies= changes systemd: send status updates to the init system systemd: update sd-daemon.[ch] systemd: properly format status message systemd: enable status messages by default daemon: make sure we never choke on SIGPIPE daemon: return successful error code when we ran successfully daemon: reset signals on initialization systemd: send status message when shutting down daemon: unblock all signals by default systemd: update sd-daemon.[ch] from upśtream init: fix lfs init script generation main: fix minor indenting mishap i18n: run make update-po client: update libavahi-client so that it can deal with auto-activated avahi daemons build-sys: bump soname Michael Biebl (1): build-sys: fix build for systems that have gtk2 but no gtk3 Piotr Drąg (1): l10n: Updated Polish (pl) translation to 100% Yuri Chornoivan (1): l10n: Updated Ukrainian (uk) translation to 100% 甘 露 (1): l10n: Updated Chinese (China) (zh_CN) translation to 100% Lennart -- Lennart Poettering - Red Hat, Inc. ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
[avahi] [ANNOUNCE] Avahi 0.6.26
Heya, just rolled an Avahi update: http://avahi.org/download/avahi-0.6.26.tar.gz This is mostly a bugfix release but also fixes a low risk security issue and adds a couple of minor new features. * Fix CVE-2010-2244 (Ludwig Nussel) * Support for Gtk+ 3 and Gtk+ Introspection * Native systemd socket activation support * Add systemd service files * Add various resource control options, for traffic rate limiting as well as cache size and D-Bus client object limits. * i18n updates * Minor other updates Aaron Faanes (1): Corrected some spellings of 'privilege' Alan Knowles (1): gobject: implement gobject introspection Andre Klapper (1): Compile with -DGSEAL_ENABLE. Fixes http://avahi.org/ticket/305 Cheng-Chia Tseng (1): l10n: Updates to Chinese (Taiwan) translation Chris Couzens (2): python: extend ServiceTypeDatabase.py API a little service-types-db: add 6 new service types Jake Li (5): l10n: Updates to Chinese (China) (zh_CN) translation l10n: Updates to Chinese (China) (zh_CN) translation l10n: Updates to Chinese (China) (zh_CN) translation l10n: Updates to Chinese (China) (zh_CN) translation l10n: Updates to Chinese (China) (zh_CN) translation James Laska (1): init: fix fedora init script handling for PIDFILE (rhbz#600430) Jorge González (2): l10n: Updates to Spanish (Castilian) (es) translation l10n: Updates to Spanish (Castilian) (es) translation Julien Blache (1): core: fix build on kFreeBSD Kees Cook (1): avahi-publish: allow A-record-only publication in avahi-publish Lennart Poettering (63): freebsd: compatibility fix for looking for CMSG_DATA fix minor typos init: Fix mandriva initscript to properly quote potentially empty variables when testing init: suse, status reports dead instead of unused when stopped fix various compiler warnings buildsys: various buildsystem updates core: Bug fixes for cleanup code systemd: add support for acquiring socket from systemd from now on we enforce a strict whitespace regime build-sys: enable git commit hook by default build-sys: make things more quiet build-sys: almost completely make the build clean, missing are howl and avahi-gobject get rid of a lot of old svn cruft discover: don't show python backtrace when we cannot connect to X11 (rhbz#599137) discover: don't show python backtrace when we cannot connect to Avahi (rhbz#585061) service-type-db: add Window Shifter protocol (See #308) gtk: build for both gtk2 and gtk3 darwin: don't make plists executable (See #301) build-sys: fix reference to LIBINTL/INTLLIBS docs: document what to do on AVAHI_ERR_DISCONNECTED (see #278) build-sys: set MONO_SHARED_DIR= to avoid build env troubles on redhat (stolen from our rpms) build-sys: fix compiler warnings to make build more silent avahi-autopid: set metric for default route in action script daemons: drop unnecessary setsid() calls daemon: make internal limits configurable daemon: add some minimal OOM checking against D-Bus service-type-db: add libvirt service limits: considerably bump up limits, since users can now easily lower them again and some applications need the high values daemon: don't include long prefix blurb in ssh-sftp.service socket: ignore EINTR on sendmsg() watch: ignore EINTR in the main loop man: clarify avahi-daemon.conf.5 dbus: move interfaces into standard D-Bus interfaces directory build-sys: get rid of old cruft dnsconfd: add --syslog switch build-sys: install systemd service files i18n: update LINGUAS i18n: run make update-po i18n: update gitignore daemon: bump up resource limit of open file descriptors a little autoipd: fix debian's dhcp path core: always normalize configure host/domain name static-services: unescape host name when using it in static service names avahi-ui: fix wording of the description strings of a few properties libdns_sd: implement DNSServiceQueryRecord() iface: check for IFA_F_DEPRECATED core: don't check ARCOUNT to avoid incompatibility with EDNS0 dbus: allow root to set the host name unconditionally core: properly unescape host name before using it in the workstation service avahi-browse: replace unprintable chars when printing services names i18n: make avahi-discover translatable. i18n: run make update-po core: don't warn if we try to remove a querier that is already gone core: fix counting of workstation services that are in the process of being registered main: fix minor memory leak main: don't abort if daemon_close_all() cannot close valgrind fds fedora: init LSB header to init scripts core: implement packet rate limiting and enable it by default
Re: [avahi] Announce Service for another Computer
On Fri, 14.05.10 16:03, Jens Dede (i...@jdede.de) wrote: Is it possible to announce a service for another computer which name is not resolvable via DNS or mDNS? I like to use python with dbus. bus = dbus.SystemBus() server = dbus.Interface( bus.get_object(avahi.DBUS_NAME, avahi.DBUS_PATH_SERVER), avahi.DBUS_INTERFACE_SERVER) g = dbus.Interface( bus.get_object(avahi.DBUS_NAME, server.EntryGroupNew()), avahi.DBUS_INTERFACE_ENTRY_GROUP) g.AddService(avahi.IF_UNSPEC, avahi.PROTO_INET6, dbus.UInt32(0), self.name, self.stype, self.domain, self.host, dbus.UInt16(self.port), self.text) g.commit() If I set self.host to an IP(v6)-Address or an unresolvable name I get an error message. self.host = uses my hostname to publish the service. How can I announce a service for any address / host? You must specifiy a proper hostname when registering a service. That means you might need to register both a host name and a service if you want to publish a remote service on a machine that lacks any kind of DNS setup. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Getting org.freedesktop.Avahi.TimeoutError in avahi browser with client-publish-service.c example
On Thu, 06.05.10 09:20, Victor M. Palacio Tárrega (vpala...@intemotechnologies.com) wrote: Hello All, I'm developing an embedded application with mDNS. Using example from client-publish-service.c works fine, but, after some time I get this with the services browser instead of TXT data: *Error: org.freedesktop.Avahi.TimeoutError: Timeout reached* I'm having the same problem with every service that is present in /etc/avahi/services. This sounds as if mDNS packets are not properly delivered. This might be because you run a firewall that drops packets somewhere or your driver does not do multicasting properly. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Remotely configure IP
On Wed, 21.04.10 15:08, Carlos Castro (carloscast...@gmail.com) wrote: Hi everyone! I want to remotely change a machine IP, so I can configure it exactly the way I want. So I'm using avahi to use Zeroconf and publish this machine on the network. Now that I can find it, I wonder if there is some kind of software ou command, so I can configure it's IP address from my workstation... I know I could use telnet ou ssh to connect to this machine and change from there, but I want to know if such software exists. Nope. Avahi cannot do that for you. Changing IP adresses remotely is problematic due to security concerns, so this should probably done only via an authenticated protocol, and no protocol Avahi implements is that. IIRC DHCP has a command for updating client configuration for active leases. You might want to investigate that. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] program keeps crashing on avahi_entry_group_add_service_strlst()
On Mon, 08.03.10 12:19, ecforu (ecforus...@gmail.com) wrote: ret = avahi_entry_group_add_service_strlst(group, AVAHI_IF_UNSPEC, AVAHI_PROTO_UNSPEC, 0, serviceName, regType, domain, host, port, txt); So, where is the code where you initialize the txt string list? When I step through the last call in my program is avahi_entry_group_add_service_strlst. Then I get the following stack dump from dbus. I have also see this happen from the avahi_entry_group_new call above. But if it makes it passed this one, it dies on the next one with the same dbus stack trace. Looks as if you are passing an invalid txt string list there. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] General Avahi usage questions
On Wed, 17.02.10 15:18, Linus Lüssing (linus.luess...@web.de) wrote: Another question that comes in my mind here: What happens, if two computers have two different subnets configured which can't reach each other over this one but could reach each other over their ipv6 link local address as being on the same ethernet link. Shouldn't avahi announce this link local address as well so that those two hosts could communicate with each other anyway? Is there a reason why avahi is not announcing all ip addresses available on an interface (with adding its prefered priorities so that the another host could choose the interface with the highest priority stated and reachable as well)? We want to minimize the traffic generated and hence pick only one address per iface to announce, and that's the one that is probably the most usefil one, i.e. a routable address. Anf if the routes mentioned above are set up properly then things should be quire robust. Ah, that's a good point, yes. Makes sense to not announce all the ip-addresses then. Is there a way to not announce an IPv4 (or Ipv6) address on an interface to save some more traffic? I found those use-ipv4/6 options in the config, but they seem to be for the protocol being used fot the announcments, not the ip addresses being announced (unless I did something wrong here which might have led me to a wrong conclusion). There's limited control for that with publish--on-ipv4= and publish-a-on-ipv6=. And I guess, avahi is announcing both the best ipv6 and ipv4 address letting the other hosts decide which type they are capable connecting to, right? Not sure I follow. Avahi will pick the best address per interface and per protocol and announce those. On a host with one network interface and both IPv6 and IPv4 it will hence announce exactly two addresses. Or let me ask it differently. The other host is getting an annoucement of both an ipv4 and ipv6 address. Which one of those should the application then choose? Should it try the IPv6 one first and try again the Ipv4 address after a timeout? When getting the announcement over IPv4 (or IPv6) only, but with both an IPv4 and IPv6 address in it, should the application then try to connect to the announced IPv4 (or IPv6) address first as that was the protocol over which it got the announcement and is having the more likely chances to succeed in connecting? Or should an application always try to use the IPv6 one first, as the connections should be a lot more stable because there are a lot less address collisions in IPv6 compared to IPv4 and therefore a lot less reassignments (if any) of IP addresses? If you use getaddrinfo() you can leave the protocol to use unspecified and then glibc will figure something out for you. It is highly recommended to do things that way, and not add any IPv6-specific hacks to applications. Traditionally glibc will then first ask for IPv6 and then fallback to Ipv4. With more recent glibc nss-mdns could actually ask for both addresses at the same time, but I never found the time to actually implement that in nss-mdns, so even with newer glibc you effectively end up asking for ipv6 first and for ipv4 second. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] General Avahi usage questions
On Sat, 06.02.10 18:46, Linus Lüssing (linus.luess...@web.de) wrote: hi guys. I was looking for a way to get pidgin/libpurple working with avahi+ipv6. but then I realised, that I'm lacking some basic knowledge about avahi :) first question: I think I remember having the avahi-daemon running which assigned an ipv4 address on a new alias interface. when does this happen, when not? avahi-autoipd is usually used as a fallback for dhcp. I also found this avahi-ipauto thingy, but it does not seem to be running for this avahi-ip4ll stuff to assign a 169.254.0.0/16 adress -169.254.0.0/24 and 169.254.254.0/24 (if I remember right) uh? avahi-autoipd will assign one of those addresses if no other address is configured via dhcp or so. 2nd question: how can I look up the ip addresses that are assigned to mylaptop.local (mylaptop should just be the local host name in /etc/hosts on 127.0.1.1 on a debian system, right?) getent hosts mylaptop.local (you need nss-mdns installed for that) Avahi will always announce the best address it can find on each interface. Meaning that global addresses are generally preferred over link-local ones. 3rd question: when exactly is avahi-dnsconfd needed and why? Is my assupmtion right, that I need this thing, if I want to use applications that are not utilising avahi (for instance a normal ping tool) and I want to reach mylaptop.local with that? It's useful when unicast-DNS server adresses are stored in multicast DNS. it could be used to configure unicast DNS servers in a non-dhcp environment. It is a more exotic feature and hence not installed by default on most distros. 4th question: so an application that uses the libavahi can tell the avahi-daemon to announce a certain service with the specified port. which ip addresses would/should the avahi-daemon be announcing then? as mentioned above avahi uses the best ip adress of each respective iface, and treat every iface seperately. 5th question: what has to be kept in mind when using avahi with ipv4+ipv6 or ipv6 only? read this article in the german linux maganzine 03/06. at the end it is saying that for ipv6 stateless autoconfiguration, avahi-dnsconfd is necessary. sounds plausible because of the lack of a dhcp server. but why is this not the case for avahi with ipv4? Some awesome dude wrote that article. He is really brainy and smart. I am a great admirer of his ;-) That said, I cannot remember I wrote necessary in that article. useful, yes. But necessary, no. In IPv4 if you do automatic configuration of routable IP adresses you use DHCP. And DHCP allows configuration of unicast DNS servers anyway. That's why the avahi-dnsconfd stuff is not really that useful there. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] network down does not remove services
On Thu, 14.01.10 12:32, Trent Lloyd (lath...@bur.st) wrote: there's a potential improve to avahi here, a pre network interface down mode to send goodbytes. Based on clairvoyance, so that we can now before the user unplugs or downs the iface that he will do so? Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] FW: avahi in ad-hoc network
On Thu, 07.01.10 09:46, ecfu (ecforus...@gmail.com) wrote: I have switched to BATMAN advanced which runs on layer 2. However I am still not seeing the avahi services across the hop. If all my nodes are within range of each other, I do see all the services. But as soon as I move one out of range so that it needs to use another node as a hop, I lose the services even though the node that is providing the hop sees the services from all the other nodes. Its weird because with BATMAN advanced, it says it runs at layer 2 so your ethernet communications should be transparent. Does this make sense that it still is not working? Nah. Link-local multicasting uses special addresses on ethernet. Maybe batman doesn't handle them properly? You might want to contact the batman folks about that as they probably know quite well if they support mcast properly. I don't really have any gateways in my network. But I suppose the hop node works as one. However in an ad-hoc network there is no way to tell if you are going to be the hop or not. What would the consequences of enabling enable-reflector=y on all nodes be in an ad-hoc network? The reflector stuff will reflect on mdns level and cannot deal with network graphs that are cyclic. I am pretty sure batman networks are cyclic. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] auto-ipd and multiple addresses
On Thu, 12.11.09 10:33, James, Neil (neil.ja...@gdcanada.com) wrote: We would like to use auto-ipd to assign IP addresses to slave DSP process (we don't want to port auto-ipd to the DSP). Is this possible? No, not really at this time. Will we have to modify auto-ipd to achieve this? Yes, I assume so, since avahi-autopid currentl refuses to do anything if there already is another Ip configured on the iface. The information contained in this e-mail message is PRIVATE. It may contain confidential information and may be legally privileged. It is intended for the exclusive use of the addressee(s). If you are not the intended recipient, you are hereby notified that any dissemination, distribution or reproduction of this communication is strictly prohibited. If the intended recipient(s) cannot be reached or if a transmission problem has occurred, please notify the sender immediately by return e-mail and destroy all copies of this message. Dude, if this really matters to you, then please dont post to this ML. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] (no subject)
On Mon, 04.01.10 14:13, Joshua Foster (jfos...@dtiweb.net) wrote: Please provide us with a minimal test case for your code so that we can reproduce your issue. Also, please do not send HTML emails to this ML and do set a subject. This message may contain sensitive, proprietary and/or privileged information. Disclosure is not authorized unless provided in writing by Defense Technologies, Inc. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose, or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. You probably should not send emails to this mailing list (which archives every mail you send to it) if you take this legal blurb seriously... Thanks, Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi-daemon crashing with r6040 device driver
On Mon, 04.01.10 17:19, Mike Miller (mmiller...@gmail.com) wrote: Hello, I'm having a problem with the avahi-daemon that I'm hoping someone can help me with or possibly point me in the right direction. I'm trying to run it on a PC-104 single board computer (SBC). Everything works fine on Debian etch, but on anything higher (lenny, sid) I'm getting a kernel panic. It's only happening with the r6040 driver that comes with the SBC for its on-board network card. The avahi-daemon runs fine with the asix driver on a second ethernet adapter that's connected via USB. [ 160.781618] BUG: unable to handle kernel NULL pointer dereference at 0004 This is a kernel problem, some bug in the multicast handling in your network driver. Please bring this to the attention of the kernel/driver developers! Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi in ad-hoc network
On Tue, 29.12.09 23:07, Conner, Eric (eric.con...@qinetiq-na.com) wrote: I have avahi running on 3 nodes in an ad-hoc network running over B.A.T.M.A.N. I have a center node that provides a hop between 2 end nodes. I can ping each of the node from any of the other nodes as well as ssh into each of them as well so I know the network is well established over the BATMAN daemon. The center node can see the published avahi services of each end but the 2 end nodes can not see the avahi services of the other ends. Is there some special configuration needed to get avahi to browse or publish over the hop? Has anyone ever configured avahi over an ad-hoc network that requires a link routing protocol like BATMAN? I am not sure on which level BATMAN operates. If it works on the ethernet level and provides correct mcast routing for link-local mcast groups then Avahi should work just fine. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] [PATCH] autoipd: also use dev:avahi when using ifconfig in linux action script
On Fri, 19.06.09 15:40, Peter Korsgaard (jac...@sunsite.dk) wrote: avahi-autoipd.action.linux creates dev:3 aliases when using ifconfig instead of the dev:avahi aliases used when ip is used. Fix it to always use dev:avahi for consistency. Thanks! Applied. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Setting a timeout value for .service file
On Thu, 13.08.09 16:03, Petri Virkkala (petri.virkk...@ee.tamk.fi) wrote: Hi all, I'm working on software for mesh networking that uses Avahi for service discovery. I'm publishing the mesh stations in /etc/avahi/services/ with my own service type. The .service file is created when a user joins the network and deleted when user leaves from it. The problem is that if user is disconnected from the network by - let's say unplugging the usb wlan adapter - the station still appears in the Avahi Discovery (and in the software) for too long. Is there a way to control the timeout when using .service files? Uh, dynamically dropping .service files into /etc/avahi/services/ is really hackish, as Robin already pointed out. The TTL can only be set in the lowest-level DNS RR registration C API. If you need to fiddle with the TTL than you need to use avahi_entry_group_add_record(). However, I am not sure that playing games with the TTL like this is really a good idea anyway. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] how to browse only two (or all) service types
On Fri, 31.07.09 09:39, Ali Mazhar (ali_...@hotmail.com) wrote: Hi Trent, Thanks for the info. My application is for a mesh network and therefore I need to see all the workstations which are in the same network just like on line users in a messenger! You shouldn't be using _workstation for this. This service used to be registered by some older MacOS versions and Avahi versions, too. But we don't register it anymore by default. And it's not exactly clear what this service stands for but it certainly doesn't stand for what I used to think it stands for and I doubt it stands for what you think it stands for either. Btw, how is it possible to remove a published service, e.g. multicast, from avahi without restarting the daemon. Actually, my application simply publishes a multicast stream port using avahi-publish and then start VLC application but when VLC is closed the multicast stops but service still remains in avahi daemon unless I restart it which naturally brings down my avahi browse application. Any tips? As Trent mentioned, just kill avahi-publish. Or avahi_entry_group_free() if you use the C API. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Multicast DNS and the Unicast .local Domain
On Mon, 13.07.09 10:14, Carsten Strotmann (private) (cars...@strotmann.de) wrote: Recent BIND 9 implementations create certain virtual zones even if no zone of that name is specified in the DNS Servers configuration. These zones are for example the zones for the loopback and RFC 1918 private IP Address reverse zones. When queried for the SOA of these zones, BIND returns a valid SOA with a serial number of 0. It might be a solution to configure a .local unicast domain that is created to stop .local queries to leak out to the Internet with a serial of '0' and have the Avahi startup script to check against this SOA serial value. A .local unicast domain used locally in a LAN would have a serial number != 0. So if Avahi detects a .local unicast zone, it can check against the serial number. If serial is 0, Ahavi can start without creating an issue for the .local DNS lookups. If the serial is != 0, Avahi can stop with an warning message to not conflict with a .local unicast zone in use. Sounds like a good idea. Could you please bring this to the attention of the Debian/Ubuntu folks who ship that .local detection script? (To my knowledge the other distros still don't, though they should) Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] /purple-2/libchatdalesa.so is not loadable: undefined symbol: avahi_glib_allocator
On Mon, 29.06.09 01:04, evanescent (coolnu...@gmail.com) wrote: Hi when it load my piding plugin it returns the above error. what will possibly I hv missed with. the plugin get compiled with no problem. Taht symbols is in libavahi-glib. You need to include avahi-glib in your pkgconfig line. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] undefined symbol: bonjour_buddy_add_to_purple
On Mon, 29.06.09 14:47, evanescent (coolnu...@gmail.com) wrote: when my purple plugin is loaded the above error is returned. what will be the possible reason ? That symbol is unrelated to Avahi. You need to figure out yourself where it is defined and that you link properly to that library. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] foo = foo.local search path
On Sun, 28.06.09 09:23, Bill Moseley (mose...@hank.org) wrote: What I'd like is to be able to say ssh foo and have the resolver first search /etc/hosts (files in nsswitch.conf), then search for foo.local, and then if that fails search for foo.hank.org via a normal dns lookup. We used to support that. But I removed it because that cannot work properly. The simple reason is that would make every name a potential mDNS name which would then mean that we'd have to resolve every hostname via mDNS first, which is not really useful however, since resolving host names that don't exist with mDNS will result in long timeout. Which hence means, resolving *any* name will take 2s or so. I guess you need to accept that typing .local as suffix for all local hostnames is the only viable option. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Resolving many ( 200 ) items seems to lead to apparent congestion in the dbus
On Mon, 29.06.09 11:05, Daniel Wynne (daniel.wy...@mobotix.com) wrote: Hi Trent! These are great news for us! :-) We were seriously concerned about the issue, that the Avahi mDNS implementation prevented us from developing Linux camera management applications for huge setups. In the following I provide a short calculation to give you a hint of what we need: - Right now we have setups with deployments of up to 800 devices We do not think that deployments with even more devices in one network make sense, as this already can be considered as a individual feasibility study. But lets calculate with this worst case scenario and for the sake of convenience we even round it up to N = 1000 devices per network. - Since every device typically comes up with 2-3 ip addresses, we have to reserve cache entries for N*( 3 A's, 1 PTR, 1 SRV, 1 TXT. ) = AVAHI_CACHE_ENTRIES_MAX = 6*N = 6000 entries. - As the dbus limitations are concerned the maximum number of dbus objects per client will have to be increased, too. Since we have to create N*3 AddressRecordBrowsers, N*1 ServiceResolver, 1 Client, 2 ServiceBrowsers = OBJECTS_PER_CLIENT_MAX = 4*N + 3 = 4003 ~ 4000 entries As I do not have that much insight into the Avahi-code there might be even more limits that might have to be adapted to achieve a coherent configuration. If you need any additional information, let me know. Thank you Guys for the great support so far! Hmm, with allowing more simulatenous clients the memory usage of the cache will increase too. I wonder how bad that actually wuld be. I.e. Daniel, how does the VSZ of the avahi daemon change when connected to a en empty network vs. one of your fully equipped ones? Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] foo = foo.local search path
On Mon, 29.06.09 17:19, Bill Moseley (mose...@hank.org) wrote: On Mon, Jun 29, 2009 at 2:24 PM, Lennart Poetteringlenn...@poettering.net wrote: We used to support that. But I removed it because that cannot work properly. The simple reason is that would make every name a potential mDNS name which would then mean that we'd have to resolve every hostname via mDNS first, which is not really useful however, since resolving host names that don't exist with mDNS will result in long timeout. Which hence means, resolving *any* name will take 2s or so. Ok. But, can you explain why mDNS would be first? Wouldn't the typical search order be for /etc/hosts, then DNS, and mDNS? Sure you could put mDNS last. But then still for *every* non-existing lookup you'll get a 2s timeout while traditionally you get a quick NXDOMAIN reply. Also note that many providers/companies/AP providers do weird things with unknown domains in their DNS servers, i.e. redirect it to their search engine, yadda, yadda. So if you put mDNS last mDNS won't be reachable at all anymore, on those networks. Also, this way you leak quite a bit of information on your local network to the internet, since every local lookup would first go to the internet DNS servers. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Multicast DNS and the Unicast .local Domain
On Fri, 19.06.09 19:19, Carsten Strotmann (cars...@strotmann.de) wrote: Hi, I stumbled over the topic I describe below when I updated an Ubuntu System from Version 8.04 tro 9.04. Avahi refused to start because I have a unicast .local domain in my network(s). This behavior is documented as recommended for distributions in the Avahi Wiki at http://avahi.org/wiki/AvahiAndUnicastDotLocal I think this is a not well thought out decision. It would be a good decision if it would detect a used unicast .local domain, but in my case, the .local domain is one of many pseudo domains that are configured as empty DNS zones on all resolving DNS Servers on the network edge (border to the Internet), to prevent any pseudo TLD like .local to be leaked into the Internet and hitting the Root DNS Server System. This is simply broken. Your DNS server should not return a working SOA for .local. If at all your DNS server should always return NXDOMAIN for all .local names, overriding the internet .local SOA. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Disappearing and reappearing service if a text is changed
On Thu, 18.06.09 09:05, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: So avahi (often) sent IP from eth1 instead of br0. Again, is the IP-Address of br0 not sent if avahi takes IP from eth1? Avahi treats every interface with an IP address assigned independently. I am not sure why you run avahi-autoipd on your bridge member ifaces in the first place. Avahi distuingishes interfaces by their ifindex, not by their addresses. Not sure what ifindex is, you mean such one like eth0, eth1 etc.? No. I mean the ifindex, as returned by if_nametoindex(). Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Disappearing and reappearing service if a text is changed
On Wed, 17.06.09 18:07, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Simple question: What should I do, so that Avahi sends IP of br0 (192.168.7.120)? The source address for UDP packets is chosen automatically by the kernel as part of the routing table logic. One question about that: Is such UDP packet sent only one time? Or twice if a device contains 2 addresses (eth1/br AND eth1:avah)? There's no such thing as eth1:avah. That's just a compat kludge for legacy tools like ifconfig. As mentioned, use ip addr insetad of ifconfig and it will show as a normal labeled address assigned to that interface. Avahi distuingishes interfaces by their ifindex, not by their addresses. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi and the loopback interface
On Thu, 04.06.09 15:09, ZioPRoTo (Saverio Proto) (ziopr...@gmail.com) wrote: Hello, I'm the author of the OLSR mDNS plugin. (http://www.olsr.org). Please look here for details about the mDNS plugin: http://zioproto.ninux.org/wordpress/2009/04/04/olsrd-mdns-plugin/ When I was testing my software I tried to configure the plugin to decapsulate the mDNS traffic to the loopback interface, so that the router itself is able to see mDNS traffic. However it seems that avahi does not work by default with the loopback interface, so I tried to following patch: http://avahi.org/attachment/ticket/198/avahi-core.patch Now if I launch the avahi-discover I see also the loopback interface listed, but I don't see the services announced by other machines in the network. With tcpdump I can see that my application is sending mDNS messages on the loopback interface, so I was expecting them in the list when launching avahi-discover. If I have an application sending mDNS packets to the loopback interface, are this packets going to be processed by Avahi using the patch? There is any other patch I can try ?? Not to my knowledge. mDNS on the loopback is a messy thing actually, since enabling it by default might cause services to show up multiple times for normal applications. Also, multicast support on loopback interfaces is supported differently on different systems, so we should probably shortcut packet delivery and bypass the stack for it. In other words: what you want to do doesn't work and is unlikely to work ever with Avahi. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Resolving many ( 200 ) items seems to lead to apparent congestion in the dbus
On Thu, 28.05.09 13:43, Daniel Wynne (daniel.wy...@mobotix.com) wrote: But I think thats exactly the problem in our case. In our Testnetwork reside about 250 cameras we want to find and resolve via Avahi. So the cache is way big enough. No. The limit of 500 entries in the cache that Avahi enforces is for RRs, not services or hosts. For each service you need (at least) one A, one PTR one SRV and one TXT RR. Usually even more. I.e. if you register 250 services and want to browser for them all at the same time you need as absolute minimum space for 1000 entries, and practically even more. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Disappearing and reappearing service if a text is changed
On Thu, 28.05.09 13:21, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Ok, this means the host always(?) get 169.254.7.XXX-Address before getting IP from /etc/network/interfaces, right? Thank you for your tip, but I cannot use ip addr on client if I want to know the IP of the host. Uh? Not sure I understand what you mean. Avahi only announces one address per host name and interface. It tries to be smart and picks the routable address over the IPv4LL address if both are configured on the same interface. Other mDNS implementations my have implemented this differently, and might even register both addresseson the network. When Avahi discovers that it is not deterministc which one of those it picks. On a local network it shouldn't matter which adress is picked for connection as long as it is a valid one. That's why it is not deterministic which address is handed to applications when they try to resolve a host name. Sorry, you're saying 169.254.7.XXX-Adress is a valid address? Here its not valid since the network-setting is given (static IP, netmask ...). If I try to ping from client, I get Destination Net Unreachable. Yes, of course. 169.254.x.x addresses are valid ipv4ll adresses. Some older OSes/distros/stacks don't know them. However everything recent knows it. In fact Windows knows them since 95b, MacOS X since about forever. And the better Linux distributions since quite some time too. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Disappearing and reappearing service if a text is changed
On Wed, 27.05.09 15:42, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: avahithread.c(203): service REMOVE: service='Floh' type='_boxMySQL._tcp' domain='local' avahithread.c(195): NEW: service='Floh' type='_boxMySQL._tcp' domain='local' Seems like the servcie refresh packets didn't get through. So its not my fault (programming side)? Dunno. I don't know your code. The timeout is reached because we didn't get a reply in time. You probably have some kind of packet loss on your network (or in the local queuing). I don't think I'm loosing packets often here, but I'll test it. Can I increase the timeout-value? Sure. Check for the TIMEOUT_MSEC value in avahi-core/resolve-host-name.c. mDNS is not a reliable protocol, by design. You should not make reliability assumptions it cannot provide. Sure... but is there any alternative for keeping a list of hosts and its IPs? I think not. Uh? Dunno what you want to do. But what you just said here sounds like a pretty generic problem, which could be solves by DNS too, heck, even a /etc/hosts file. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Disappearing and reappearing service if a text is changed
On Wed, 27.05.09 15:47, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Hi, I found out, after timeout avahi tells me the host does have such address like 169.254.9.181 which is wrong (true: 192.168.0.120). Where the hell does avahi get such address? And why does it happen? That's an IPv4LL address. It gets assigned by avahi-autoipd or a similar tool. Most modern systems use it as fallback when no manual configuration and no DHCP is available. Umm... on Server the IP is set manually to 192.168.0.120 so it has a fixed address. This is why I'm wondering about the IPv4LL address displayed on client. How can I get rid of this? Sometime I just got 169.254.9.XXX. (this is bad) Sometime I got 169.254.9.XXX and after some seconds I got 192.168.0.120. (this is ok) It's not the client side that makes these adresses up. It's the host itself which configures those addresses. On Linux use ip addr to find out which addresses are assigned to your interface. On a local network it shouldn't matter which adress is picked for connection as long as it is a valid one. That's why it is not deterministic which address is handed to applications when they try to resolve a host name. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Resolving many ( 200 ) items seems to lead to apparent congestion in the dbus
On Wed, 27.05.09 09:49, Daniel Wynne (daniel.wy...@mobotix.com) wrote: Though its not the standard case, we deal with more than 250 hosts for testing purpose. We want to resolve all of them, either without spamming the DBus-System or crash into any Avahi-limitation. First, we tried resolving everything right on the way, but this lead to apparent DBus-congestion, which could only be solved by restarting the whole service. After that we tried a queued system which allows only a few resolvers to be coexistent, but this did not lead to any perceptible improvement. The DBus still seems to be congested after a short while. D-Bus congestion? I don't think that exists. What exactly makes you think D-Bus could be 'congested'? So this was a misunderstanding, sorry. In a previous thread somebody mentioned that every D-Bus client can have at maximum 250 objects. I assumed that every resolver browser is a client's object. If this is correct, the limitation is forced by Avahi and not the D-Bus system. Yes. That is true. For security reasons Avahi enforces limits on all resources a local or remote client can allocate and control. This might have to do something with the internal limits Avahi applies on almost everything: in this case possibly the size of the chache? Also note that if you issue a lot of requests the local IP stack packet queueing might already drop packets. Lost packets will most likely result in timeouts. Is there an easy way to verify this? Could not find any proper logfile. Edit avahi-core.c and the function avahi_cache_update(). Look for these line: if (c-n_entries = AVAHI_CACHE_ENTRIES_MAX) return; Then add a log line there that prints something when an entry is not added to the cache because it reached its max size. i.e. if (c-n_entries = AVAHI_CACHE_ENTRIES_MAX) { avahi_log_debug(cache: overrun); return; } or something like that. Then run avahi-daemon --debug to get a peek on the debug output. The max cache size is controlled via AVAHI_CACHE_ENTRIES_MAX in avahi-core/cache. It is set to 500. Given the number of hosts this might actually be way to low for your use case. Try to increase it to 5000. This solution is not really applicable since it requires recompiling the sources. It's free software. Sources available. Is there a way to cleanly free the browser resolvers immediately? No. But this 'recycling' of resolves/browsers shouldn't hurt in your case. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Disappearing and reappearing service if a text is changed
On Mon, 25.05.09 15:57, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Hi, I wrote two applications (1 for server and 1 for client). On Server-Side a service-file (in /etc/avahi/services/...) is written if service-text is changed. Sometimes on Client I got such DEBUG-Info: avahithread.c(203): service REMOVE: service='Floh' type='_boxMySQL._tcp' domain='local' avahithread.c(195): NEW: service='Floh' type='_boxMySQL._tcp' domain='local' Seems like the servcie refresh packets didn't get through. Which means, the service disappeared and reappeared which is really ugly for clients. :( It would help me if you could tell me why this happens (and maybe how it can be avoided). And seldom such bad thing happens (which must not happen :( ): (Resolver) Failed to resolve service 'Floh' of type '_boxMySQL._tcp' in domain 'local': Timeout reached Is there a way to get out why Timeout is reached? The timeout is reached because we didn't get a reply in time. You probably have some kind of packet loss on your network (or in the local queuing). mDNS is not a reliable protocol, by design. You should not make reliability assumptions it cannot provide. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Disappearing and reappearing service if a text is changed
On Tue, 26.05.09 12:12, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Florian Erfurth schrieb: And seldom such bad thing happens (which must not happen :( ): (Resolver) Failed to resolve service 'Floh' of type '_boxMySQL._tcp' in domain 'local': Timeout reached Hi, I found out, after timeout avahi tells me the host does have such address like 169.254.9.181 which is wrong (true: 192.168.0.120). Where the hell does avahi get such address? And why does it happen? That's an IPv4LL address. It gets assigned by avahi-autoipd or a similar tool. Most modern systems use it as fallback when no manual configuration and no DHCP is available. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] About how avahi-dnsconfd adds the hosts
On Mon, 18.05.09 14:19, Félix Arreola Rodríguez (fgatuno@gmail.com) wrote: Hi Well, this is still some new for me, but since I discover how avahi works, and his easy use I was wondering with too many things. Avahi resolves everything. Well, to the point, when a host (for example, named gatuno) starts, the other hosts with avahi adds with the name gatuno.local. Can I change this?, for example, that name the of the host be gatuno, without the .local. .local is the simply mDNS domain. WHile in theory you could manage other domains with it it realistically wouldn't make sense. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Resolving many ( 200 ) items seems to lead to apparent congestion in the dbus
On Tue, 19.05.09 08:10, Daniel Wynne (daniel.wy...@mobotix.com) wrote: Hi Lennart! Though its not the standard case, we deal with more than 250 hosts for testing purpose. We want to resolve all of them, either without spamming the DBus-System or crash into any Avahi-limitation. First, we tried resolving everything right on the way, but this lead to apparent DBus-congestion, which could only be solved by restarting the whole service. After that we tried a queued system which allows only a few resolvers to be coexistent, but this did not lead to any perceptible improvement. The DBus still seems to be congested after a short while. D-Bus congestion? I don't think that exists. What exactly makes you think D-Bus could be 'congested'? Only a restart of the service solves the problem. Without deep insight I would assume, that the DBus-objects created are not freed-up cleanly or lets say immediately. They seem to remain and so the bus gets filled-up. To give you a more detailed overview I explain to you what exactly happens: - We browse the .local domain for HTTP and HTTPS services with two coexistent browsers - For every resolver we create, we create another one to receive additional address records - The first resolver is freed immediately after resolving, the second after the last address has been resolved - This works fins for about 100 Hosts, then it starts to stumble. The other resolvers achieve timeouts. If a timeout occurs we trie to resolve again. - After about 1min we receive another circa 50 hosts before the next stumbling - Then about every 1-2 minutes we receive a variable amount of hosts We also tried a queued algorithm with only a certain amount of coexistent resolvers without a noticeable improvement. This might have to do something with the internal limits Avahi applies on almost everything: in this case possibly the size of the chache? Also note that if you issue a lot of requests the local IP stack packet queueing might already drop packets. Lost packets will most likely result in timeouts. As you have a now a litlle more detailed view on what we do, could you please give us some hints on how to improve the use of Avahi? Dunno. This really depends on the problem. I don't have such a setup here, so I cannot debug this myself. If you get timeouts then either the query or the response packets got lost. Try wireshark to find that out. Maybe its not an DBus but an Avahi issue? mDNS/DNS-SD is not a reliable protocol. It wasn't designed to be one. Don't make assumptions on its reliabilty that you shouldn't make. What are the exact limits of Avahi concerning cache, timeouts queries, etc...? The max cache size is controlled via AVAHI_CACHE_ENTRIES_MAX in avahi-core/cache. It is set to 500. Given the number of hosts this might actually be way to low for your use case. Try to increase it to 5000. How does the free-mechanism work exactly, especially on the DBus-side? Resolvers/browsers are ref counted by the D-Bus clients. I.e. if two clients issue the same queries they end up with the same internal browser/resolver. Also, a browser/resolver stub will be kept around after the browser resolver died until the next time we'd have to reissue a request for it in the hope we might be able to reuse it later on for another client. That allows us to resuse the same browser/resolver for repeated queries of the same data. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi scalability again
On Fri, 15.05.09 14:11, Olivier R-D (o@laposte.net) wrote: Heya, I am wondering about using avahi (for service discovery ) for an application, on one lan, but with potentially many computers (maybe 1000 ) and especially many services to publish (maybe 1). Have some of you tested avahi scalability at that level ? mDNS has not been designed for such networks. Apple did some scalability tests with mDNS. And IIRC they say it should work up 200 hosts or so. I never did similar tests for Avahi, but I'd assume that Avahi isn't too far off, given that we pass the Bonjour compat test. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] corrupted double-linked list
On Fri, 08.05.09 09:11, Gaspard Bucher (gasp...@teti.ch) wrote: I cannot seem to get avahi to work in my application. During runtime I get *** glibc detected *** ... corrupted double-linked list This seems to be related to avahi in very strange ways. Has anyone experienced such problems when using avahi ? Any help would be greatly appreciated. I really feel stupid here... code: http://github.com/rubyk/rubyk/tree/master/oscit/src/linux Uh. I don't really grok that code. Bu uh, you seem not to lock the threaded main loop ever. Which gives me the idea that this issue is simply a race issue, because of incomplete locking. Please note that Avahi is not thread-safe by itself. It offers you a few tools to use it in a threaded environment however. But you have to do all locking explicitly yourself, Avahi won't do it for you implicitly. Basically the idea is that after you started the threaded event loop if you need to touch the avahi context in anyway from anothe thread you have to lock the event loop first (i.e. temporarily interrupt its execution) and then unlock after your access. i.e. something like this: snip /* First lock things */ avahi_threaded_poll_lock(ml); /* The call Avahi client functions */ foo = avahi_client_xxx(bar); waldo = avahi_service_browser_yyy(humpf); /* When you are done, unlock things again */ avahi_threaded_poll_unlock(ml); /snip This locking logic works very similar to Java monitors if you know those. Or it's even a bit like the MT support in the GLib event loop. Hope this helps, Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Resolve multiple IP4 Addresses
On Thu, 07.05.09 17:29, Mark Gollahon (mgolla...@exacq.com) wrote: On Thu, 2009-05-07 at 19:57:37 +0200, Lennart Poettering wrote: On Thu, 07.05.09 13:09, Mark Gollahon (mgolla...@exacq.com) wrote: In my case, we are connecting to Axis cameras. Some models of Axis cameras report their configured IP address first; others report their link-local address first. Now, because of certain other design limitations and requirements that would take too much time to go into here, we have to store the camera's IP address when that camera is configured in our system. So, if during our camera detection phase, Avahi hands back the link-local address, then we're totally sunk when the link-local address changes (which does at the most inopportune time - when the camera reboots and we're trying to reconnect). I am sorry, but if you use IP addresses like this it is not more than an ugly hack. Avahi, as it is written, doesn't give me much choice in the matter! In my scenario, I can rely on a configured IP address much more than on a link-local address or even a name resolve. The link-local has a short lifespan (at most between reboots) and I cannot depend on someone entering name-to-ip maps into the hosts file. Therefore, it makes sense to store and connect to the cameras using an IP address and not the name. So, you call this an ugly hack. I do not think so, but if it is a hack, it is a realistic one. I am not exactly sure I fully understand what you really want to do. But I kind of get the idea that you should make your device publish a service that includes some real identifcation cookie. Then, use that cookie for id and nothing else. Even if you do not use IPv4LL addreses, what else is there? DHCP? That isn't much better, unless someone configured it to hand out static addresses. My whole point is that IP addresses are not suitable for identifying devices. If you want to identify devices use some other mechanism, publish a service with some unique ID (MAC address?) or so. IP addresses are not useful for identifying machines. IP adress can change. IP addresses can be assigned to multiple machines. One machine, even one network interface can have multiple IP addresses. IP addresses are only unique and suitable for identification in a very specific context, which is when you talk about network interfaces for the use of routing on a specific network in a specific time frame. I agree that this is a good argument for *not* using IP addresses to refer to devices on a network. However, it is also a great argument *for* returning multiple addresses during an mDNS query. If my device has multiple addresses (i.e. link-local and a configured one), why not return all to the querying process and give it the ability to *choose* which way to contact the device? All of the name resolving mechanisms I know about have the ability to return multiple addresses to the program using it, all except Avahi. Why would you want to know those addresses? Those addresses should be used for creating connections to a device right after you resolved the address. They cannot be assumed to be stable outside the current network or the immediate time frame. On the local network all IP addresses of a device can be equally used to communicate with it. If you need a routable, stable IP address, then mDNS/DNS-SD is *not* where you should be looking. The way mDNS/DNS-SD is designed is that browsing, resolving, connecting happens within the same location and timespan. If you change location or time then you should browse and resolve again. You seem to want to simply skip those steps then and go directly to connecting which may work sometimes but not in the general case. Furthermore, trying to explain what that 169.x.x.x address is or why a user doesn't see the address they configured into the camera to a user not versed in IP addressing is a complete non-starter. Yes, IP addresses should generally not exposed to the user. That's why host names have been invented. Yes, I very much agree with this statement. I would much rather deal with hostnames than IP addresses. However, hostnames can change, too, so its not the panacea you are making it out to be. Sure, they can change. In many cases if someone changes the identity of a device like that then he probably has a reason for that and you shouldn't try to outsmart him. In other cases neither IP address nor the hotname are suitable, but some other kind of ID. D-Bus' machine id would be a good choice (/var/lib/dbus/machine-id). Therefore, saying that scenarios where an mDNS recipient needs all addresses for a given DNS name is a misuse of mDNS really is disingenuous. There really are times when we need all of the reported addresses we can get. No. This is a hack. An ugly one. And a misuse of mDNS/DNS-SD. This is precisely the reason why I chose not mimic the Bonjour API
Re: [avahi] Resolve multiple IP4 Addresses
On Thu, 07.05.09 16:09, Daniel Wynne (daniel.wy...@mobotix.com) wrote: ok... ...this is the call I make in the ServiceResolverCallback: ... avahi_record_browser_new ( client, AVAHI_IF_UNSPEC, You should specify the same interface you discovered via the service browser/resolver. AVAHI_PROTO_UNSPEC, You should specify the same protocol you discovered via the service browser/resolver. name, You should specify the same host name you found via the service resolver. AVAHI_DNS_CLASS_IN, AVAHI_DNS_TYPE_HINFO, HINFO? I thought you were interested in the address records (i.e. A or ). AVAHI_LOOKUP_USE_MULTICAST, Leave that to the implementation, pass 0 here. The Callback of the RecordBrowser always goes into state CACHE EXHAUSTED. http://avahi.org/download/doxygen/defs_8h.html#f7ff3b95259b3441a282b87d82eebd87 Hey, read the docs, that's why I wrote them. They might not be complete, but they do explain what CACHE EXHAUSTED means. You'll get a couple of callbacks. If you don't care about CACHE EXHAUSTED then ignore it. In which format does name has to be? A little more documentation would be helpful ;-) Read the DNS-SD spec. This is an API to lookup low-level RRs. The parameters are hence in the format those low-level RRs use. i.e. fqdn. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Resolve multiple IP4 Addresses
On Thu, 07.05.09 13:09, Mark Gollahon (mgolla...@exacq.com) wrote: In my case, we are connecting to Axis cameras. Some models of Axis cameras report their configured IP address first; others report their link-local address first. Now, because of certain other design limitations and requirements that would take too much time to go into here, we have to store the camera's IP address when that camera is configured in our system. So, if during our camera detection phase, Avahi hands back the link-local address, then we're totally sunk when the link-local address changes (which does at the most inopportune time - when the camera reboots and we're trying to reconnect). I am sorry, but if you use IP addresses like this it is not more than an ugly hack. IP addresses are not useful for identifying machines. IP adress can change. IP addresses can be assigned to multiple machines. One machine, even one network interface can have multiple IP addresses. IP addresses are only unique and suitable for identification in a very specific context, which is when you talk about network interfaces for the use of routing on a specific network in a specific time frame. Furthermore, trying to explain what that 169.x.x.x address is or why a user doesn't see the address they configured into the camera to a user not versed in IP addressing is a complete non-starter. Yes, IP addresses should generally not exposed to the user. That's why host names have been invented. Therefore, saying that scenarios where an mDNS recipient needs all addresses for a given DNS name is a misuse of mDNS really is disingenuous. There really are times when we need all of the reported addresses we can get. No. This is a hack. An ugly one. And a misuse of mDNS/DNS-SD. This is precisely the reason why I chose not mimic the Bonjour API in this aspect: people start to misuse the technology for things it wasn't designed for. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Avahi-autoipd Help.
On Tue, 21.04.09 11:46, Andres E. Rodriguez Lazo (andres...@gmail.com) wrote: Hello all, I've installed avahi-autoipd only on a Fedora 11 VM. I'm trying to assign an interface called eth1 an IP address thourgh avahi-autoipd command, and it tries to bind the address but it keeps conflicting. Do you know why this is happening and it cant assign an address? This is what it shows to me: Dunno. Might have something to do with the bridging logic of your VM solution. Try wireshark to find out if there actually are conflicting ARP packets sent out. [r...@localhost ~]# avahi-autoipd eth1 Found user 'avahi-autoipd' (UID 499) and group 'avahi-autoipd' (GID 499). Successfully called chroot(). Successfully dropped root privileges. Starting with address 169.254.2.176 Callout BIND, address 169.254.2.176 on interface eth1 Received conflicting normal ARP packet. Callout CONFLICT, address 169.254.2.176 on interface eth1 Trying address 169.254.189.32 Callout BIND, address 169.254.189.32 on interface eth1 Received conflicting normal ARP packet. Callout CONFLICT, address 169.254.189.32 on interface eth1 Trying address 169.254.39.206 Callout BIND, address 169.254.39.206 on interface eth1 Received conflicting normal ARP packet. Callout CONFLICT, address 169.254.39.206 on interface eth1 Trying address 169.254.66.131 Callout BIND, address 169.254.66.131 on interface eth1 Received conflicting normal ARP packet. Callout CONFLICT, address 169.254.66.131 on interface eth1 Trying address 169.254.223.251 And so on. Thanks for yor help ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] keep avahi-browse-list uptodate
On Thu, 16.04.09 16:25, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Hi, Question A: if somewhere a host is (un)plugged from/to network (without shutdown or stopping network). It takes very long till the host (and its services) disappears from avahi-browser. I can resolve the problem by doing following: /etc/init.d/avahi-daemon restart Is this the proper way to refresh avahi-browser-list? No, certainly not. Question B: Is there a way to reduce the Timeout, till a host will be removed from avahi-browser-list? (same example like Question A if I (un)plug a remote Host) Kind of. The registrants can control this by setting the TTL of the records they register. In avahi_entry_group_add_record() this its the 'ttl' parameter. That function can be used to register 'raw' RRs. For the higher level calls such as avahi_entry_group_add_service() you don't have this fine-grained control. It is not recommended to play games with the TTL however. You have to balance out the traffic imposed on the network with the delay it takes until services disappear from the network when the registering machine died without sending goodbye packets. Never forget: mDNS is not a reliable protocol. And we don't want fool anyone into believing it was. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] [ANNOUNCE] Avahi 0.6.25
On Tue, 14.04.09 07:04, William Immendorf (will.immend...@gmail.com) wrote: On Mon, Apr 13, 2009 at 9:32 PM, Lennart Poettering lenn...@poettering.net wrote: William Immendorf (1): add init script for LFS WAIT! You forgot the lfs intscript fixes I sent to this list earlier. It is in one of my earier emails. Please make a rerelease (0.9.25-1 will be perfict) with my patch in it! Nah. Not sure this matters enough. Please prepare a patch according to my comments to that patch I just mailed. I'll then apply it to my tree. You distro should ship that patch on top of the upstream packages until we do the 0.6.26 release. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] [PATCH] Fixes for lfs initscripts
On Tue, 14.04.09 14:57, William Immendorf (will.immend...@gmail.com) wrote: On Tue, Apr 14, 2009 at 2:31 PM, Lennart Poettering lenn...@poettering.net wrote: On Mon, 06.04.09 09:41, William Immendorf (will.immend...@gmail.com) wrote: 1. I didn't use a sed to redict the .in file to a bootscript because there was nothing to sed. However, this causes chown to fail because it cant find the 'avahi' bootscript. I fixed that in the patch by adding a cat to redict the .in file to the bootscript itself. Uh. It would make more sense if the init script would start the daemon via the right path. Echg, you don't know how loadproc from the LFS initscripts works: It automaticaly finds the daemon to start, so seding is not necessary. Fine. Most real world distro's tend to prefer absolute paths to avoid the needless stat()s. Anyway, if no rewriten is necessary, then please remove the rewriting entirely. No 'cat foo.in foo' please. That's nonsense. 2. And also, I forgot to make symlinks to the bootscript in the makefile, because that is what LFS does. So I added the ln commands to make the symlinks. We dn't create those link farms on any distribution. Activating init services is a job for the administrator, not for 'make install'. You don't know how LFS initscripts are layed out. The main init script is installed into /etc/rc.d/init.d, and symbolic links are created in /etc/rc.d/rc{0,1,2,3,4,5,6,sysinit}.d dirs. The form they are in is this: {Either S for start, or K for kill}{number that decinisates when the script will run}{bootscript name} Yes, that's classic SysV init. Like almost every distro still does it. And we don't create those link farms on any distro. Oh, and why is this in make install? Because it would be a hassale for the sysadmin to create those symbolic links manualy. Sure. But that's why folks invented package managers. We don't create those symlinks for any distro because creating those links is a job for a package manager, not for install scripts. Also, creating those links manually is almost certainly wrong. Every distro has a tool that manages that automatically for you. Also, I thought the point of LFS was to do everything manually, wasn't it? Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] [ANNOUNCE] Avahi 0.6.25
On Tue, 14.04.09 14:59, William Immendorf (will.immend...@gmail.com) wrote: On Tue, Apr 14, 2009 at 2:35 PM, Lennart Poettering lenn...@poettering.net wrote: On Tue, 14.04.09 07:04, William Immendorf (will.immend...@gmail.com) wrote: Nah. Not sure this matters enough. Please prepare a patch according to my comments to that patch I just mailed. I'll then apply it to my tree. You distro should ship that patch on top of the upstream packages until we do the 0.6.26 release. Echg, my patch WORKS and is good and should be added to 0.6.25 ASAP. That something works is not necessarily an indication that it's ready to be merged. I am sorry, but doing releases is a lot of work. And this triviality is not reason enough for me to do another release right-away. Sorry. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
[avahi] [ANNOUNCE] Avahi 0.6.25
Heya! http://avahi.org/download/avahi-0.6.25.tar.gz Yes, I suck big time and still haven't merged Sunil's LLMNR work. But it's going to be in .26. Promised! This is mostly a release just before the F11 freeze. Contains the low-risk security fix CVE-2009-0758. Alexandru Szasz (1): Added Romanian translation Daniel Nylander (1): Updated Swedish translation Diego 'Flameeyes' Pettenò (2): Ignore macro files added by libtool 2.2. Remove gettext macro, the package uses intltool instead. Diego Elio 'Flameeyes' Pettenò (1): Don't enable PIC code for all compilation. Domingo Becker (2): updated Spanish translation a minor change Fabian Affolter (1): Updated German translation Lennart Poettering (11): Merge branches 'master' and 'master' of ssh://rootserver/home/lennart/git/public/avahi include sys/types.h since libcap needs that apparently Use send_destination for DBus deny rule run make update-po update LINGUAS run make update-po Make .desktop files pass desktop-file-validate hide more files from git bump version and soname update NEWS update README Piotr Drąg (1): Updated Polish translation Rob Leslie (1): CVE-2009-0758: Reflector creates packet storm on legacy unicast traffic Sharuzzaman Ahmat Raslan (1): Updated Malay translation Taylon Silmer Lacerda Silva (1): Updated Brazilian Portuguese translation Timo Jyrinki (2): Updated Finnish translation against current git trunk. Update Finnish translation against current trunk. All strings translated. William Immendorf (1): add init script for LFS Xavier Conde (1): Updated catalan po kelemeng (1): Sending translation for Hungarian kmilos (2): Sending translation for Serbian Sending translation for po/s...@latin.po Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] 1 avahi_simple_poll_loop for 2 (or more) servicetype?
On Mon, 06.04.09 11:02, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Lennart Poettering schrieb: On Wed, 01.04.09 16:07, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Now I've got a problem. I would like fire a function, if a host is not in network anymore (e.g. somebody pulled out ethernet-cable). Same on client (on which my small avahi-application runs) if I pulled out the network-cable. If I disable ethernet, it works fine! :) Idea? You'll get a BROWSER_REMOVE callback when a service goes away or the network interfaces it was found on becomes unavailable. I tested it again with AVAHI_BROWSER_REMOVE, but unfortunatelly it takes long time if I pull out the cable. If I turn off eth0 then it works fine. Is there no way to detect connection lost (by pulling out the cable) as soon as possible? I'm afraid the answer would be like: No. :( Read about use-iff-running in the avahi-daemon.conf(5) man page! Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] 1 avahi_simple_poll_loop for 2 (or more) servicetype?
On Tue, 31.03.09 15:13, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Hi, I successfully wrote a small tool which get strings (txt-record) from a Host and keep uptodate (if Host changed Text, Client will get new text). For this purpose I use type _boxMySQL._tcp. The host does have other service-type too, called _boxAVG._tcp. I want keep a list of Hosts with such service and add/remove Host to/from list (with AVAHI_BROWSER_NEW/ AVAHI_BROWSER_REMOVE). Now I've got a question. Is this possible to do both in same poll_loop? I've tried avahi_service_browser_new() with _box*._tcp, but it didn't work. :( Do I need 2 different poll_loop (one for _boxMySQL._tcp and other for _boxAVG._tcp)? You can add as many browsers/resolvers to a single event loop object as you wish. Use one browser per service type. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] 1 avahi_simple_poll_loop for 2 (or more) servicetype?
On Wed, 01.04.09 16:07, Florian Erfurth (f.erfu...@reel-multimedia.com) wrote: Hi, after examinating the source browse-avahi.c I found some answers myself. I can use only 1 simple_poll by using 2 different browser (the one for service_type and the other for service). It works great. Now I've got a problem. I would like fire a function, if a host is not in network anymore (e.g. somebody pulled out ethernet-cable). Same on client (on which my small avahi-application runs) if I pulled out the network-cable. If I disable ethernet, it works fine! :) Idea? You'll get a BROWSER_REMOVE callback when a service goes away or the network interfaces it was found on becomes unavailable. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi w/fixed IP, ipv6, and Macs?
On Mon, 30.03.09 16:36, David Abrahams (d...@boostpro.com) wrote: Hi, I have a Linux fileserver called hydra on my network with a fixed IP address. I can ssh hydra.local from another Linux box, but not from a Mac. Adding an /etc/avahi/hosts file containing 192.168.xxx.xxx hydra.local doesn't help. If I pass -vvv to ssh, it appears that the Linux boxen connect to hydra's ipv4 address, but the Mac is finding the ipv6 address. I don't even know how the ipv6 address is getting assigned --- I don't set up any ipv6 in hydra's /etc/network/interfaces file --- but it seems as though this should all still work. I realize I could turn off ipv6 on the fileserver, but I shouldn't have to, should I? Avahi will return the first address found. If your client app (i.e. ssh) didn't specify the proto it would prefer it is not defined whether IPv4 or IPV6 is returned. You can try connecting with ssh -4. Or simply configure your sshd to listen on the Ipv6 interface if you have configured it. IPv6 is configured automatically by modern kernels unless expclitly disabled. You can also teach Avahi on the server side to not use Ipv6 (set use-ipv6= to No -- which is the default except on Debian where the maintainers decided to set this to yes -- against my explicit requests and publish--on-ipv4=no) Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] browsing for services on a specific multicast address
On Tue, 17.03.09 09:24, Alon Noy (alon@ilsemedia.nl) wrote: OK, that makes sense, so I need to browse for the subtype right?! So here is the next question, is this a valid subtype: 225_0_10_100._ftp._tcp? No. It needs a _sub as second label. Also using an IP address in the type simply makes no sense to me... I think that avahi insists on having a leading underscore but is it an avahi think or is it in the zeroconf spec? http://files.dns-sd.org/draft-cheshire-dnsext-dns-sd.txt Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] browsing for services on a specific multicast address
On Wed, 18.03.09 18:59, Alon Noy (alon@ilsemedia.nl) wrote: I guess I didn't quit understand what the apple util dns-sd suppose to do (I'm not a maci :) ), and the partial (and misleading) documentation which I received for this server farm didn't help either, but I think that what is really going on is that I'm suppose to query for sub type: _225_0_10_100._sub._ftp._tcp in order to identify the nodes of a cluster. the sub type _225_0_10_100 is just a way to group the different servers in a cluster. Its a way to say that these servers communicate between themselves on multicasting address 225.0.10.100 and not as I initially thought that they are using a different multicasting address for mDNS. I am not sure if I now get it, but this still smells like a misuse of mDNS/DNS-SD to me. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] Removing an individual service from an entry group
On Thu, 05.02.09 09:28, Jim Hodapp (james.hod...@gmail.com) wrote: I'm writing my own little avahi-daemon since my application is an embedded target. I've successfully used the avahi-core and avahi- common libraries to publish a service. I've used the function avahi_server_add_service() to add the service. If I want to individually remove this service (or unpublish it), how do I accomplish this? I do not seem to see a function like avahi_server_remove_service() or the equivalent. I've tried just resetting the entry group, but that seems to put things into a weird state if I want my avahi-daemon clone to republish a new service in the future. The point of entry groups is that all entries change state simultaneously -- and that includes deleting them. If you have entries whose state is manipulated independantly then you should have multiple entry groups. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi
Re: [avahi] avahi-browser returns IPv6 address rarely instead of ipv4
On Fri, 30.01.09 09:53, sundararaj reel (sundararaj.r...@googlemail.com) wrote: Note that if your application only supports ipv4 .. your -application- can request to publish on v4 only. (and browse on v4 only) without disabling v6 system wide. How is it done (browse on v4 only)? I could not find an option to avahi-browse that restricts browsing to ipv4 addresses only and man page of avahi-browse does not talk about ipv4 . Or am I using the wrong application to browse for services ? I am using avahi-browse 0.6.20. avahi-browse doesn't allow you to do that. The underlying API does, but that's not exposed in this specific tool. Patches welcome. Lennart -- Lennart PoetteringRed Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 ___ avahi mailing list avahi@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/avahi