UKenGB

I am configuring snomONE with a Snom M9 and several softphones and I'm after some ideas about how best to integrate a DECT system into the whole setup. In particular Intercom between extensions. Any DECT system will have some method of calling other extensions and snomONE has its Intercom feature, but the latter only works between handsets NOT registered to the same extension. So in theory, I could have ALL the M9 handsets on one extension and they could use snomONE Intercom to call any of the softphones, but NOT any of the other M9 handsets. But the M9 handsets could use their own internal calling to other handsets. However, although that maybe covers all requirements, it needs 2 different methods to make the Intercom call (I assume) so you need to know where the recipient is before calling, which may well defeat the object of calling in the first place. So has anyone gone down this route? What would be the best way to integrate the 2 requirements so that Intercom is possible between any 'clients' (i.e. M9 handset, or softphone), but using the same method of calling so users only need to know ONE way to call another client? What about trying to call ALL the other clients if you don't know where the intended recipient is. How have others achieved this?

I don't believe that to be the case. I could consistently run it on the Mac Pro and it would consistently fail on the Mac Mini. They are BOTH running OSX 10.6.4 (which is Darwin 10.4) with the latest updates. The difference in the OS is that the Pro is running the regular client version and the Mini is running the server version and although there are undoubtedly differences, I very much doubt there exists such a fundamental difference in their respective implementations of these POSIX calls. I realise it might be considered immaterial what the problem was now that it does run correctly, but I believe it is always worth knowing exactly what caused it so that similar problems can be avoided in the future. Hence I think we need to understand fully what was occurring here.

OK downloaded the new binary and it's up and running without any bother. Thanks for a very prompt response to the problem. But what WAS the problem? As you say, there are Macs running the version that refused to run on my Mini. If it's that gettime function, why did it only affect some Macs and not others? One thing I have noticed however is that from System Profiler, my OS is Darwin 10.4.0 which is significantly different to yours. Are you guys actually testing and running this on Snow Leopard? What version of OSX are you using to test this?

In which case, why does it run OK on some Macs with the same OS?

There's also still the question of the user who should start the pbx. The working directory has been created all owned by root, so if anyone else tries to start it, pbx crashes because it cannot write to anything. IMO this ought to generate a user readable error so they can see what to do (without having to start it again in debug mode etc), but that's a minor point. However, what is the user that SHOULD start pbx? Should we create a user just for pbx (like a lot of service processes) or just run as root? Either way, the docs should really point this out.

It does look like both Paul's and my problem is related to some gettime function, but why would getting a slightly incorrect elapsed time cause it to crash and why is it trying to do this when it's just trying to start up and why does it not affect ALL OSX installations? In truth though, those questions are irrelevant to those of us who just need it fixed so we can get pbx running. Presumably this means a new build with suitable corrections. When might we see that?

How can we best progress this? Shall we just continue here or set up a direct dialogue outside the forum? BTW, are there any known instances of snomONE running OK on OSX Snow Leopard SERVER? I can't see any specific references from any posts here saying it does. Is that perhaps the problem? How tested was it on the Server edition of OSX? Was that the Mini Server? XServe? Other Mac running SLS?

If that were the case I'd still recommend better formatting, but not only do things never work that well, but in this case we are instructed to modify the ports if they might clash with those already in use, before running the software, by editing this file. It's a small point (and I have bigger issues as you are aware pbxnsip), but every other user of XML files that I've come across sees fit to format them. Anyway, at least we know it doesn't cause a crash:-)

That just produces the not entirely helpful: Starting up Bus error BTW, what user should be able to start it? Should any admin user be able to start it, or does it have to be root? Or who? This seems to be important. This is just so disappointing. I am trying to test this on what seems to me to be a very simple and obvious setup. It's a Mac Mini Server (Snow Leopard of course) with the latest updates. It is running several services like DNS, DHCP, eMail, AFP (not HTTP currently) for light use in a home environment and is exactly the setup I want to be able to deploy snomone. Now maybe I'm being unrealistic, but a product such as snomone that advertises itself for use on Macs ought to run on this setup OUT OF THE BOX. I know this has been raised elsewhere, but why on earth configure it to try and use the standard ports that will DEFINITELY clash with the services that any Mac server is likely to be running. I guess I wouldn't mind so much if it just needed some tweeking to get it running well, but when it simply crashes immediately with no real indication of what the problem is, it's kinda annoying. Although the error doesn't sound like it's a port clash problem, I changed all these ports (SIP/HTTP/HTTPS), but still the same error. Next?

Well there are some settings only available within that file. Even the docs say you may have to edit it, so can I suggest that some proper formatting would make a lot of sense? As you say, it's easy enough to modify the format, except that when you (try to) start the pbx, it immediately re-writes the file as one single line again:-( I realise that there should be no need for formatting, but all XML editors I've tried baulk at this lack of formatting. However, it is obvious that this is not the reason why the pbx is so totally unable to run on my Mac Mini Server. Back to the drawing board:-(

I've now also installed it on a MacPro and unless I run it with sudo, it also crashes for apparently the same reason. But with sudo, it seems to run fine:-) Ah ha, so back to the Mini and run it with sudo - and it crashes. I delete all the snomone files from the Mini and install it again. Then I run it with sudo - and it crashes:-( Anyone else had problems like this?

Just one long single line. Looks to me like someone did a bad job of bringing this over from DOS/Windows:-( Is it supposed to be like this, or should there be some line breaks somewhere? Presumably the pbx app can read it whether the line breaks are DOS, Mac or Unix?

I've already posted in the Snom forums that I had a problem with the install since it is supposed to finish with the PBX actually running, but it doesn't. I think I now know why. The app crashes as soon as it tries to run. When I try and start it, I get the messages to the screen saying it's starting (but that's just from the startup script) and then nothing. If I check with ps -ef | grep pbx, there's no process running. On further investigation I discover it's simply crashed. The crash log is below, but this is very disappointing. This is on a Mac Mini Server with Snow Leopard Server and all the latest updates. It couldn't be more standard. So what's snomONE having so much trouble with? Is there a later update that fixes this? Process: pbxctrl-darwin9.0-2011-4.2.0.3950 [45848] Path: /Library/snomone/pbxctrl-darwin9.0-2011-4.2.0.3950 Identifier: pbxctrl-darwin9.0-2011-4.2.0.3950 Version: ??? (???) Code Type: X86 (Native) Parent Process: ??? [1] Date/Time: 2010-10-22 15:31:05.485 +0100 OS Version: Mac OS X Server 10.6.4 (10F569) Report Version: 6 Exception Type: EXC_BAD_ACCESS (SIGBUS) Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000000 Crashed Thread: 2 Thread 0: Dispatch queue: com.apple.main-thread 0 pbxctrl-darwin9.0 0x003c27cb clock_gettime(int, timespec*) + 525 1 pbxctrl-darwin9.0 0x003c27f9 clock_gettime(int, timespec*) + 571 2 pbxctrl-darwin9.0 0x003002c7 g726::output_limiting(int) + 91505 3 pbxctrl-darwin9.0 0x002e7312 dtmf_recognition::count_history(char) const + 22336 4 libSystem.B.dylib 0x925b61bf __cxa_finalize + 208 5 libSystem.B.dylib 0x925b60d4 exit + 33 6 pbxctrl-darwin9.0 0x00053fb9 std::map<int, int, std::less<int>, std::allocator<std::pair<int const, int> > >::operator[](int const&) + 2763 7 pbxctrl-darwin9.0 0x00054497 std::map<int, int, std::less<int>, std::allocator<std::pair<int const, int> > >::operator[](int const&) + 4009 8 pbxctrl-darwin9.0 0x001ff0a8 std::map<unsigned int, int, std::less<unsigned int>, std::allocator<std::pair<unsigned int const, int> > >::operator[](unsigned int const&) + 15672 9 pbxctrl-darwin9.0 0x00207cb3 std::map<unsigned int, int, std::less<unsigned int>, std::allocator<std::pair<unsigned int const, int> > >::operator[](unsigned int const&) + 51523 10 pbxctrl-darwin9.0 0x00001816 0x1000 + 2070 Thread 1: 0 libSystem.B.dylib 0x925970fa mach_msg_trap + 10 1 libSystem.B.dylib 0x92597867 mach_msg + 68 2 libSystem.B.dylib 0x92598c32 thread_policy + 199 3 libSystem.B.dylib 0x9260afe9 pthread_setschedparam + 219 4 pbxctrl-darwin9.0 0x001fba04 std::map<unsigned int, int, std::less<unsigned int>, std::allocator<std::pair<unsigned int const, int> > >::operator[](unsigned int const&) + 1684 5 libSystem.B.dylib 0x925c481d _pthread_start + 345 6 libSystem.B.dylib 0x925c46a2 thread_start + 34 Thread 2 Crashed: 0 pbxctrl-darwin9.0 0x003c2910 clock_gettime(int, timespec*) + 850 1 pbxctrl-darwin9.0 0x001fbb88 std::map<unsigned int, int, std::less<unsigned int>, std::allocator<std::pair<unsigned int const, int> > >::operator[](unsigned int const&) + 2072 2 libSystem.B.dylib 0x925c481d _pthread_start + 345 3 libSystem.B.dylib 0x925c46a2 thread_start + 34 Thread 3: 0 libSystem.B.dylib 0x925c4726 _pthread_start + 98 1 libSystem.B.dylib 0x925c46a2 thread_start + 34 Thread 4: 0 libSystem.B.dylib 0x925c4680 thread_start + 0 Thread 5: 0 libSystem.B.dylib 0x925c4680 thread_start + 0 Thread 6: 0 libSystem.B.dylib 0x925c4680 thread_start + 0 Thread 7: 0 libSystem.B.dylib 0x925c4680 thread_start + 0 Thread 8: 0 libSystem.B.dylib 0x925c4680 thread_start + 0 Thread 9: 0 libSystem.B.dylib 0x925c4680 thread_start + 0 Thread 2 crashed with X86 Thread State (32-bit): eax: 0x00000000 ebx: 0x003c2847 ecx: 0x00000001 edx: 0x00000000 edi: 0x00000000 esi: 0xb0103000 ebp: 0xb0102f68 esp: 0xb0102dd0 ss: 0x0000001f efl: 0x00010293 eip: 0x003c2910 cs: 0x00000017 ds: 0x0000001f es: 0x0000001f fs: 0x0000001f gs: 0x00000037 cr2: 0x00000000 Binary Images: 0x1000 - 0x4ccfff +pbxctrl-darwin9.0 ??? (???) <5753FCA1-4AC8-1E80-C36A-72ADD4B525B0> /Library/snomone/pbxctrl-darwin9.0 0x8fe00000 - 0x8fe4162b dyld 132.1 (???) <A4F6ADCC-6448-37B4-ED6C-ABB2CD06F448> /usr/lib/dyld 0x92596000 - 0x9273cfeb libSystem.B.dylib 125.2.0 (compatibility 1.0.0) <3441F338-2218-6D36-3F95-3A16FBF6713D> /usr/lib/libSystem.B.dylib 0x93b0d000 - 0x93b10fe7 libmathCommon.A.dylib 315.0.0 (compatibility 1.0.0) <1622A54F-1A98-2CBE-B6A4-2122981A500E> /usr/lib/system/libmathCommon.A.dylib 0x99853000 - 0x998bdfe7 libstdc++.6.dylib 7.9.0 (compatibility 7.0.0) <411D87F4-B7E1-44EB-F201-F8B4F9227213> /usr/lib/libstdc++.6.dylib 0xffff0000 - 0xffff1fff libSystem.B.dylib ??? (???) <3441F338-2218-6D36-3F95-3A16FBF6713D> /usr/lib/libSystem.B.dylib

Great, at least it is possible. Couple of things though:- So you're saying set up the alternative routes as dial plans with different priorities and then change the priorities to switch routes? Dial plans are stored as XML files are they not? Couldn't you therefore just swap files? Or would it not see the change until a restart? Otherwise, how could one automate the URL method? Just pondering here, but how could you get cron to kick that off? I'm thinking you could use telnet to 'deliver' the URL. That's all that is required isn't it, just hit that URL? What is the response if you do this in a browser? Does it indicate success or otherwise? I want something that is truly automated, so the dial plan is adjusted automatically and calls are directed as required with no further manual intervention.

Something that seems to me to be an obvious requirement is to be able to use times and days in a dial plan. It is quite likely that different trunks will have different costs associated with the calls (at least if they are to different providers), but that can vary at different times on different days. E.g. there may be free evening and weekend calls on a PSTN trunk, but with cheaper calling at all other times via a VOIP trunk. How can this be done? I am sure that this facility is not readily available in dial plans (although as I say, I cannot for the life of me think why it is always omitted) or is it? Failing that, is there any other way to achieve this, like some form of external control (e.g. cron and shell scripts etc) to provide different dial plans according to a schedule?