andrewgroup

Heres the scenario and our testing and workaround Inbound call to main desk Transfer to HG 110 (both extension in the HG are busy or simply dont answer) Call goes to AA950 (message say press 1 to leave a message and the call goes to mail boxes via 8102 (Client opted to leave message) If not the call goes to HG 109 This rings ext 114 but if 114 does not answer the call should go to 8103 (Client did not reach anyone mailbox) We see the missed call in the SIP log on ext 114, (but the call never goes to 8103 The main desk can send call to 950 and the remaining process works as expected. Only calls from other HG going to the 950AA are lost.. Our work around was to create an agent group 108 with the same extensions and after hearing ring redirect to 8103 Now calls sent to HG110 then go to AA950 bounce over to Agent Group 108, no answer the call correctly goes to the 8013 mailbox. did I explain that well enough to understand what is happening? We rebuilt all acocunts too. Cheers -

Agreed - we are now tracking down where in the call tree structures calls are being disconnected. I'm seeing a lot of 404 not found in the SIP traces on the LAN side of the PBX include 486 and 487.. Leaving this for tommorrow. Thanks

The Interface on the Windstream SIP trunk does not have a default route. Instead we have a persistent route for the the Registration/Proxy IP address to go out what would normally be the default route on that interface. The other WAN facing IP is a static IP in a business comcast address and has the only default route. We have IP access allows for the remote offices public facing IP addresses and internal IP ranges where remote phones are registered. I do not see a range for the ITSP SIP trunk, but problems there would be a total outage. What logging might capture the error? Thanks

According to windstream Engineering they use the following ports on SIP trunks. RTP Media udp range 16384 32767 udp range 49152 53247 Signal Protocol udp 5060 tcp 5060 udp range 1718 1720 tcp range 1718 1720 udp 2427 tcp 2427 udp 2000 tcp 2000 The RTP range of 49152 - 53247 being lower than the default, may have explained callers being incorrectly connected to other callers. Will update if the problem continues. 2-14-2011

According to a senior engineer at Windstream / Nuvox the following are the supported ranges for SIP trunks. We've made the adjustment on the PBX to match the second range given for UPD 49152-53247 and we hope this addresses the issues. This now points out the need to verify this information with all SIP providers and to match accordingly. Cheers Perhaps we could have other users post the same information in SIP Provider Forums. I'll report this so that it can be easily found.

More 411, I created an extension on win32 3.4, and set ring cell phone option after 1 second, and the extension is set to record all calls. Calling this extension internally or from an external phone to its DID number are properly forwarded to the cell phone number, but in neither case are the calls recorded. does this information help? Thanks

We are burning an extension with a forward all calls to the cell phone, how should we set this up so that it is a forked call and gets recorded?

In this case, the PBX has a public IP address and no router. We are checking with Windstream to determine what UDP ports they support on their SIP trunks. The PBX is using the default values, 49152-64512, or a 15K block, how many ports does a call consume? I also mentioned the Snom Phone Peer-Peer call completion, does this have no affect on the PBX? The PBX actually has two public facing IP's the first serves the SIP trunk, and the second serves remote Snom Phones in remote locations behind firewalls of their own. The 3rd Nic is a 50 local snom phones on the LAN.... I assume the remote phones will use their full range of UPD ports, as most firewalls will pass that upper range. Assuming this is not a Router UDP issue from the ITSP, would it seem then that something on the LOCAL lan is interfering with the UPD Port Range? If the default block of 15,000 ports is by far enough ports, I can chop off the top half as a test, and wireshark for UPD activity in that range. Would that seem a likely course of action? Comments welcome.

Seems we are not recording all or any calls that are redirected immediately from an extension to a remote cell phone. Can we not accomplish this, or must we do some trickery like redirected after ringing to record these calls. Nay comments are helpful. Thanks

We have a client with version 4.2.0.3911 (Win32) The complaint is that inbound callers from their Windstream SIP trunk on a 6MB dedicated trunk that are delivered to a main receptionist desk, and then being transferred to a hunt group of a couple extensions are inadvertently being connected to other inbound callers. We are using Snom 320 handsets, and I will check that the option JOIN in TRANSFER in the advanced behaviour settings is not set, on all phones, and I'm confident they are not. The main receptionist is not.... Assuming that to be the case, how could or would you be able to connect to inbound callers together either my accident on the phone in the way you are transferring calls, or could the PBX or phone handsets be doing this with a programming error. blind transfers, attended transfers, ringbacks etc.... and how might we construct the call trees so that users cannot make these mistakes.? Let me add to my investigations, How does the PEER to PEER call completion setting in a SNOM phone affect PBX? Shouldn't this be off, allowing only the PBX to handle call txfers or does this not matter? Any thoughts, Thanks

I like the boolean thought, but the example that we think is most common is a business with 8:00 to 17:00 MF business hours.. Autoattendent 400 takes all inbound calls from the trunk and looks at the NORMAL HOURS service flag So this AA wants to take all calls M-F 8 to 5. the Other service flags on the PBX are Lunch Mon-fri 12:00-13:00 Closed Manual Open Manual Holiday Scheduled 12/25 7/4 Holiday - Emergency Manual For unscheduled holidays or weather related closes Message Boxes on the System ar; 500 for after hours messages 501 for Lunch hour messages 502 for Holiday Messages 503 for Emergency close messages If the Normal Hours are set then forward all calls to the regular Hunt Group or business auto attendent But during the time of the normal flag being set, continue to look at the other flags and branch conditionally on them.. If lunch is set during normal work hours then forward calls to the lunch message box If the manual closed flag is set, then override the current close hours and be closed normally if Emergency close is set then override the current close time actions and forward calls to the emergency message box.. the current listing of service flags and action boxes on the same line is a first come first serve and you have to be careful about how you list the automatics vs the scheduled.. This would overcome what Matt just posted too as it allows for the appearance of a reset of a schedule flag with a manual override.

A common situation is the need to provide a client with the traditional night button. Customers use this as an ad-hoc method to close early or stay late.. With the current functionality you have to build various service flags and put them in order (manual flags first) and the following are just a few CLOSED, LUNCH, HOLIDAY, SCHEDULED. then matching mailboxes with messages or autoattendants with call routing. I think the right approach might be having he ability to nest service flags. By this I mean if this service flag is set, then check and process the SUB service flags first. I think this would overcome some problems we have that result in creating more than necessary service flags, or timing issues with scheduled service flags.

I'm confusing myself with your suggestion, When you said "not using country codes" are you referring to the field where we would enter our Country code? When you say use a prefix that corresponds to real codes, I assume you mean we create a MATRIX of CODES vs. REAL dialing necessary to reach a given country. I was hoping a REGEX expression could be create that would match the first 5 digits, and if a match was correct then drop these digits and dial the remaining digits. MATCH 5 DIGITS and then SEND from the 6th position and remaining digits. Your thoughts appreciated.

We would like to create a simple but affective Dial plan that adds a 3 or 4 digit code of our choice that will get stripped out and the remaining and valid digits are sent. preceding all calls with 55532 011529981932010 pattern to Match 55532* the replacement string would next to be a regix expresssion to strip 55532 How might that look, played with some online regex testers but my head is sore from beating it against the wall..

Tom, Long ago we created a post on optimizing a windows installation, We've used various tools like TWEAKXP to create a stripped down XP installation with minimal add-ons, then set every service to manual and a reboot of windows it will start the minimal services. Of course remove file/print and all IP services, if the NIC supports UDP/TCP checksum offloading enable that too.. You'll need to enable the install services to install PBX, or do the services changes after the installation. Being of Old School decent, PBX's are important enough to run on suitable hardware, especially one that support ECC RAM. You can read all about single bit error statistics, but the reality is most PBX's are installed in non-perfect environments and are subject to many factors that can cause single bit errors. Temp, ESD, Humidity, Shock. The better hardware platforms track and register single bit corrections it's not uncommon to see these in server hardware BIOS reports.

Tom, Sorry to me late in commenting, but your problem wreaks of having PNP devices on the LAN, perhaps the gateway router supports PNP... I'll go out on a limb and assume that your setup is not running 100% PNP on the phone configs and you have phones manually registered against the local IP of the PBX? It's very common for popular L2 switches to have a featured called ICMP snooping enabled/disabled and this prevents the switch to deliver broadcast packets to all ports. I also suggest that if you L2 switch supports ports sniffing, then use it to capture the packets going to the PBX. I also suggest getting some basic tools like PRTG to track latency to the devices during calls. Getting QOS correct is important... What brand switches are you running? We have grown to use PORT based QOS assignments for the PBX port and the PUBLIC IP port. You assign the port to a Queue level, normally 1,2,3,4 in priorities, then you also assign the DSCP values used by the phones to the same Queue. You said these problems are occurring on the local LAN, so creating a scalable test environment should be easy and establishing a repeatable experiment should net some results. If you have the stomach, expertise, or time to learn, check this tool out http://sipp.sourceforge.net/

It would be a great feature to have a email notification whenever a Manual Service Flag is flipped. The notification should include the date/time and the extension that flipped the account. We had a client where only three phones are allowed to access the flag, and the business was closed for an afternoon, and nobody knew, and nobody claims to have flipped the flag.. Any Thoughts?

I've downloaded and installed from the Droid market the program called "REMOTE WAVE" from Walter Yongtao Wang Version 1.7.7 and it automatically ties intself into the audio system and when I click Open in a WAV attachment I choose and select always use remote wave and Voila, it works perfectly. While I would hope the Droid OS addresses this, a $00.99 solution is more than acceptable.

After having just upgraded to the Sprint EVO, I've since learned the WAV format used by PBXnSIP is not natively played and it seems to be well known. Numerous posts exists about this problem. Has someone found the least interuptive cure? Trying LINDA to associate WAV files, trying meridian player, and others free versions, Happy to pay a $1 or $2 for an ad-free version... no Love yet..

Rather than letting the dial plan block the call, accept the call redirect them to the error extension, play the "BAD BOY" message and let the mailbox send an email to the designated email address and voila'. A CDR will be created that identifys the caller. Might that work? Maybe a FAKE IVR will recieve the errant calls and trigger the alerts... Being in the IT business, we have a List Server, that processes all alerts from all client devices and systems and parses the subject lines and body content with boolean if-then-else logic and can forward alerts or emails or generate new emails and sends to the designated persons.

Our Updates were failure due to a switch, weird huh. Follow our thread on PNP with CS410. The WEB interface appears to simply copy the TAR file into the /pbx folder, so I guess you can simply use scp to put the tar file in place and the next reboot will update the PBX.

Initial response from support appears to be confirming our hypothesis. Enabling Multicast in the OS, creating Multicast Routes appears not to be enough. Perhaps a MultiCast Debian package is in need of being installed to make MAC - PNP work on CS410's Oh the hours spent learning the hard way. Arg..... Now in a holding pattern for escalated paid support to help us address this problem. Cheers

could someone explain the requirements of passwords to meet the available options. Maximum Security (example, at least ? characters, ? Special characters provide list) upper case etc.... Minimum Security Allow All passwords reviewed almost all DOCs in KIWI and SUPPORT and cannot find any reference on this? This would be nice to know since you can import all accounts and being sure the passwords meet the requirements is crucial.

Update, for those that are viewing this post. Our hypothesis is, that we have upgraded a batch of CS410's from a V2 version... not likely capable of supporting multi-cast in the base OS. We deployed the CS410 using the internal DHCP server, (it gave the option66) allowing the phones to be configured and regustered. Turn on another DHCP server, the generated files still existed, and the new DHCP server had no option 66 tftp setting, but the phones had been already provisioned. Factory restart the phones, and they cannot register.. CANNOT JOIN THE MULTI-CAST traffic is in the LOG files.... Perhaps the BASE OS needs to be modified or reconfigured to support multi-cast. Looking now to verify Multicast is operational on the CS410. Really don't want to force TFTP Option 66 onto smaller businesses best suited for the CS410. Check back later. Cheers

It seems we are back to seeing PNP work when the CS410 provides DHCP, and it appears to automatically setup and provide an option 66 setting to it's own IP address. We have read and assumed that PNP is now working using the ARP cache and resolve on the LAN and based on the latest snom docs, says an HTTP process improves the security. When things are working, we appear to be seeing the http process kick off after the sip registration requests happens to the localhost... When things are not working as expected (another device providing ip address only DHCP) we do not see that http processes after the subscribe message.. We did not see any traffic out the WAN port either.. I guess this leads to need for a good PNP Diagnostic Procedures. Questions; To use option 66 or Not? What needs to be verified to assure success? (no pnp devices on the lan, etc, other multicast activity. What generated files can be deleted? What files on the the PBX retain relevant information? A basic PNP flow chart of what files are generated and when?