Problems with incoming calls....

[MJ_user]

The problem I am having is with incoming calls only, the phone rings and I can pick up, but the call is dropped after 32 seconds every time.. On the magic jack display it reads "awaiting acknowledgment" so it seems like its not fully connecting... Another problem I am having is with incoming calls also.. If I don't pick up the phone/accept the call and the caller hangs up or gets my magic jack voice mail, my phone will continue to ring long after the caller has hung up..

My magic Jack worked fine for the first 30 days and magically started doing this on the 30th day exactly... Any ideas on why this problem started?

[noetic1]

The same thing started to happem with my MJ recently. Outbound calls are normally ok except for when I receive a call and it gets dropped 32 seconds later. When I try to call the person back right after the call is dropped, my outbound call fails. I get a dial tone, dial the phone number and then the call just drops and I get a dial tone again.

I can make outbound calls again after a few minutes.

I set the majicJack.exe to High Priority and applied the latest upgrade from the MJ website but it didn't fix the issue-
http://www.magicjack.com/9/magicfix.asp

I'd also like to know if someone has a fix/workaround. Thanks!

[noetic1]

I've been doing some testing and I found something interesting.

When I call my MJ using my cellphone, the message on the MJ screen says 'Awaiting acknowledgement' even though my wife already answered the phone and we are having a conversation. Then at exactly 32 seconds into the call, the call is dropped and the MJ screen goes back to 'Ready to call' status.

So, it looks like there must be some kind of timeout setting wherein if the call status is 'Awaiting acknowledgment' the call is automatically dropped after 32 seconds.

[dnjbrown]

Yes, I am having the same issue....does anyone have a fix?

Thanks

[noetic1]

I just tried bypassing my Linksys WRT54GL wireless router and was able to successfully receive MJ calls without getting dropped 32 seconds after answering the call.

I haven't made any changes to my router so does anyone have any ideas why my MJ doesn't work when connected to my Linksys WRT54GL wireless router?

[az2008]

I just tried bypassing my Linksys WRT54GL wireless router and was able to successfully receive MJ calls without getting dropped 32 seconds after answering the call.

I haven't made any changes to my router so does anyone have any ideas why my MJ doesn't work when connected to my Linksys WRT54GL wireless router?

I'm using a WRT54GL without any problem. (I flashed it to the freeware Tomato firmware. But, I didn't have any problem before doing that.). You might try doing the factory reset to the original, factory firmware values. That would eliminate whether there's any settings you made causing the problem.

I've read some people say they have to open ports on their router. But, I didn't have to do anything to mine.

I know when MJ makes (and probably receives) a call, it uses a UDP port between 10000 and 20000. What you describe sounds like it happens when this port is opened. But, like I said, I didn't have to do anything special with my WRT54GL. It worked out of the box. So, I'd question if something changed on your router (maybe to block those ports).

Mark

[dnjbrown]

Ok all, I have a question. I want to utilize port forwarding to try and keep this 32 second problem from happening.

First, what is the application name: Do I use magicjack or mjusbsp or what?

Second, the port range. I have read online that MagicJack utilized UDP only range 5060 to 5070 but above listed 10000 to 20000.

Third, anyone know how to set up QoS just for magicjack?

Thanks,

Dave

[noetic1]

I reset my router and apart from setting the basics (i.e., SSID, Password, enable WPA2 wireless security, setting up wireless MAC filters) I didn't change anything else from the default settings.

Looking through the posts on this site, I also tried manually setting the MTU on my router and added a port forward entry. I took screenshots of my current settings.

Please let me know if there is something else I need to do because I still can't get my MJ to work with my WRT54GL. Everything was fine up until about 2 weeks ago.

Verified my IP address


I have PPPoE authentication set up on my DSL modem and I read in another post to manually set the router MTU to match the modem MTU = 1492


I also found in another post to open ports 5060 - 5070 so I set that up on my router.

[az2008]

I have PPPoE authentication set up on my DSL modem and I read in another post to manually set the router MTU to match the modem MTU = 1492

I don't think MTU would cause the problem you're having. It might cause choppy audio.

There's a defined way to determine your max allowable MTU size.

ping proxy1.phoenix.talk4free.com -f -l 1500

Start at 1500 and take it down until you don't get the fragmented message. Then add 28. (1492 is what I used when I had DSL.).

I don't know why you're using port forwarding. I didn't have it on my factory 54GL. It worked fine. I don't believe the 5070 port is being initiated by MJ. And, if port forwarding were necessary, you'd need it for ports 10000 to 20000 (where the communication port is opened when a call occurs). But, I think that's opened by the MJ application running on your computer.

If you reset your router to factory default and you still have the problem, maybe the other thing to consider is whether you're running a software firewall. Another thing to try (just because it seems to fix so many problems you wouldn't expect it to) is to try a powered USB hub.

Beyond that, I don't have any more ideas. You could install WireShark and examine what happens when a call comes in. If you're not familiar with this sort of thing, you'd need to spend a lot of time playing with it, and digging through the data figuring out how to put the pieces together.

Mark

[synchron]

Ever since I setup QoS on my Wireless Airlink 101 router, there have never been any problems with call quality from incoming calls when I'm doing heavy downloading on newsgroups at full bandwidth. Before this change, it was always a problem. I have it set to high priority on ports 5000 thru 5099 which was recommended in another thread.

Synchron

[dnjbrown]

Can you post your directions for setting up QoS? That would be awesome.

Here is what I did to fix the problem last night after reading this thread and other tech support threads.

I disabled DHCP on the CPU running MagicJack, installed a static IP on that CPU, then went into the router and enabled the DMZ for that IP. The 32 second problem looks to be solved but I am still interested in how to set up QoS.

Thanks all and I hope this helps anyone still struggling.

BTW: I am using this overseas (military).

[az2008]

Can you post your directions for setting up QoS? That would be awesome.

My experience is here:

http://www.dslreports.com/forum/r212207 ... o-firmware

Mark

[az2008]

Ever since I setup QoS on my Wireless Airlink 101 router, there have never been any problems with call quality from incoming calls when I'm doing heavy downloading on newsgroups at full bandwidth.

I don't know how QoS can affect download priority. I read a couple of threads that said it can't. There's no way for your router to tell all the sending routers how to treat the data they're sending? It just all arrives at your router when it arrives?

Before this change, it was always a problem. I have it set to high priority on ports 5000 thru 5099 which was recommended in another thread.

I haven't watched what happens on an incoming call. But, on an outgoing call (as discussed in another thread) MJ seems to open a random UDP port between 10000 and 20000. That's the port that the real voice activity seems to occur on (because, without giving that port priority, the voice quality deteriorated as I uploaded a file).

MJ seems to always open a UDP port on 5070 when it starts. That just seems to be for control information. I didn't notice any performance improvement giving it priority. (Although I'd give it priority just because it might make it more responsive to whatever the control information is, like accepting and originating calls.).

Maybe this weekend I'll look at what happens when accepting calls. I'll post an update (if someone doesn't do it before me).

Mark

[dnjbrown]

Hey az2008,

Thanks for the post. Lots of good information in there.

I am going to let the DMZ fix work for now and if I have any more problems load up the QoS.

One question, under QoS or anything else, what are you all putting for "application name"? I have it as magicJack but not sure if that is the right way to list it.

Thanks all,

[az2008]

One question, under QoS or anything else, what are you all putting for "application name"? I have it as magicJack but not sure if that is the right way to list it.

I'm not sure about other routers/firmwares. But, with Tomato the only place to put an application name is in a comments field. The other fields (ip address, port) are what do something.

Mark

[az2008]

I disabled DHCP on the CPU running MagicJack, installed a static IP on that CPU, then went into the router and enabled the DMZ for that IP. The 32 second problem looks to be solved but I am still interested in how to set up QoS.

I've never used the DMZ function. But, I think that puts the computer associated with that IP address outside the firewall. If that's true, you're likely to have your computer hacked, infected with spyware, spambots, etc.

Mark

[dnjbrown]

Hey guys,

Just want you to know I have had no issues since putting my static IP in the DMZ on my linksys. I just finished a 45 min call with no delays and no hang-ups.

Dave

[az2008]

Maybe this weekend I'll look at what happens when accepting calls. I'll post an update (if someone doesn't do it before me).

I just received some calls and it works the same way (as far as QoS is concerned) as originating calls.

Destination UDP port 5070 when MJ starts. I guess this is a control channel.

When a call is originated, or received, it opens a port to a seemingly random destination UDP port between 10000 and 20000. (Probably told which port through the 5070 control port).

So, the advice you received to give priority to ports 5000 - 5090 sounds incorrect. It sounds like it gives priority to a lot of ports MJ doesn't use. And, doesn't give priority to the port that really carries all the voice during a call.

One thing I noticed is that some DNS calls are made at the time that 10000 - 20000 port is opened. If there were a problem reaching the DNS server, or the address of whatever the DNS server was asked to resolve, that 30 second problem sounds oddly related (like a timeout). I'm not sure why bypassing the router (reset to its factory defaults) would affect that.

Mark

[synchron]

You guys r right and i stand corrected, I checked with Wireshark and port 5070 is only used when mj does its regular pinging, right after the phone is dialed, and when the conversation ends.

But my Airlink has 3 settings under QoS management: Disabled, Manual (where you add application/port entries) and Auto Smart QoS which is what is selected and improves the quality. Perhaps it prioritizes any RTP (voice) packets over regular TCP packets because my network monitoring software definitely shows that the binary newsgroup downloads are taking less than the full 6MB bandwidth of my ISP and that 80k on both send and recieve is dedicated to mj when I'm doing both at the same time. Once the phone conversation ends, the maximum 6MB is give back to file downloading.

Synchron

[JohnODon]

I am curious why I am not seeing this with my D-Link DIR-655. It has a "smart" function (called "Automatic Classification") much like your Airlink. However, when I max out my newsgroups (815K/sec) and initiate a phone call, my newsgroup stream moves only about 5K (810K) and the call gets choppy. I also tried to setup a manual rules with every combination of ports I could think of (mostly on Mark's recommendation). Nothing has made a diference. I must be doing something wrong.

Mark, when you refrer to a "destination" port, would that be representative of "remote" port on my router (shown below). Maybe I have them reversed in my rules? And just to clarify, we are talking about UDP, right? (just an fyi...I do realize that I have the rules disabled...before you ask)



Thanks,

John

[longspeak]

The problem I am having is with incoming calls only, the phone rings and I can pick up, but the call is dropped after 32 seconds every time.. On the magic jack display it reads "awaiting acknowledgment" so it seems like its not fully connecting... Another problem I am having is with incoming calls also.. If I don't pick up the phone/accept the call and the caller hangs up or gets my magic jack voice mail, my phone will continue to ring long after the caller has hung up..

My magic Jack worked fine for the first 30 days and magically started doing this on the 30th day exactly... Any ideas on why this problem started?

I am a new MJ user with these same problems and a Linksys WRT54GL router.

MJ live support finally told me that I need a router that didn't have such a 'tight' firewall!!

I flashed to Tomato, and above problems went away and all is well.

[az2008]

Mark, when you refrer to a "destination" port, would that be representative of "remote" port on my router (shown below). Maybe I have them reversed in my rules?

Yes, I think you have it reversed. 5060-5070, and 10000-20000 are the destination ports on the MJ server.

You should be able to prove that the rules are taking effect by calling the echo phone number (909 390 0003), and starting a speedtest.net test. During the upload test, count from 1 to 20. Then set the priority in your QoS rule to low, and repeat the test. MJ should become unuseable if the rule is affecting the traffic the way you think it should.

And just to clarify, we are talking about UDP, right?

Yes, both rules should be UDP.

Personally, I like to make the rules applicable to the IP addresses of two MJ servers. (Or, at least the server associated with the 10000-20000 ports). I don't like the idea of giving priority to 10k ports just because MJ uses a random port in that range. But, there's always a risk the IP address will change and my traffic will fall into the default "low" category (and be unuseable).

Mark

[az2008]

I flashed to Tomato, and above problems went away and all is well.

I wish we could figure out what the 32-second problem is. I have a GL and didn't have that problem. I only flashed to Tomato to try QoS.

It's strange that a different firmware on the same router would fix the problem. There must be a lot of people using Linksys routers. But, all of a sudden 3 people in this thread say they've experienced the 32-second problem. (That's odd.).

Mark

[az2008]

Auto Smart QoS which is what is selected and improves the quality. Perhaps it prioritizes any RTP (voice) packets over regular TCP packets because

That's true. It must be using packet inspection to identify the type of traffic. From what I've read, it's better to use rules involving MAC address, IP address, port before relying on packet inspection. Less load on the router's CPU.

But, using a rule for 10000-20000 isn't pretty either. It's likely something else will open a connection to one of those ports and get the same priority as MJ. You can figure out the IP address of the vms1 host your MJ device connects to when a call is originated or accepted, and create a rule using the IP address instead of the host. But, it's just a matter of time until MJ changes the IP address (and your MJ traffic sorts into the low-priority category).

Mark

[JohnODon]

Auto Smart QoS which is what is selected and improves the quality. Perhaps it prioritizes any RTP (voice) packets over regular TCP packets because

That's true. It must be using packet inspection to identify the type of traffic. From what I've read, it's better to use rules involving MAC address, IP address, port before relying on packet inspection. Less load on the router's CPU.

But, using a rule for 10000-20000 isn't pretty either. It's likely something else will open a connection to one of those ports and get the same priority as MJ. You can figure out the IP address of the vms1 host your MJ device connects to when a call is originated or accepted, and create a rule using the IP address instead of the host. But, it's just a matter of time until MJ changes the IP address (and your MJ traffic sorts into the low-priority category).

Mark

But in the case of my router, I can give remote ports 10000 - 20000 priority if the traffic is coming my MJ box ONLY (which has a static IP). If another box on my network uses a remote port in that range, it will not get priority because it is not define by the rule. And being that my MJ box is just that and only that, I really should not have any other rougue traffic coming from it. the last thing I will need to chec is RemotelyAnywhere.

That should solve that issue...at least in my little brain!

John

[noetic1]

I've been going back and forth trying to figure out what the heck went wrong with my MJ. I bypassed the router, switched cables, used different computers, uninstalled MJ, reinstalled MJ, opened ports on the computer and router and still had the same problem where incoming calls dropped after 32 seconds because the status was 'awaiting acknowledgement'.

I didn't want to leave my computer running the MJ in the DMZ because it would be outside the firewall.

I actually recently bought my WRT54GL router about 2 months ago to replace an old Dlink DI-624. The WRT54GL was running great up until about 2 weeks ago when incoming calls started to get dropped. So, I was convinced I had a bad MJ dongle since my router is relative new still.

In my previous post, I noted that I could receive calls if I disconnected my router. But since I need my wireless router, and I've burned so much time already, I figured I'd upgrade the firmware to dd-wrt since another poster upgraded his to Tomato and it worked.

After troubleshooting for the past 5 hours, it turns out that all I had to do was upgrade my WRT4GL firmware to the latest dd-wrt firmware. Apart from the normal wireless security settings, I left things as the default settings (i.e., I didn't set up any port forwarding) and I am now able to successfully receive calls with getting dropped.

AZ2008-
I have no idea why the out of box firmware on my GL prevented me from receiving incoming calls even with port forwading set up but the latest dd-wrt firmware fixed it for me.

Link to latest dd-wrt firmware-
http://www.dd-wrt.com/dd-wrtv3/dd-wrt/downloads.html

I would highly recommend upgrading to dd-wrt if your router is capable of being upgraded to save you all of the troubleshooting time and effort if your incoming calls are getting dropped.

I'll send a follow up note in a couple of days to verify that upgrading the firmare to dd-wrt is stil working and there are no issues.

[noetic1]

I forgot to ask. Can someone tell me how to set up QoS on dd-wrt. Since there has been so much discussion about QoS, I'd like to set it up to see if it makes a noticeable difference.

[MJ_user]

I am still having the problem as well, I even had a Magic Jack support person call me to help.. He was very nice and took a lot of time working with me. He used wire shark to determine it was the router blocking out the return, and it is the MJ that is timing out from not receiving it.. He was baffled as to why, he says he has the same router at his house and his is working fine, he checked all my P/C and router setting via FOG CREEK Copilot and everything was good, but still did not work. I am using a WRT54GS router with the most current firmware, it is just strange why it stopped working out of the blue. It does work OK if the router is set in DMZ mode, but I really don't want every port opened.. He suggested I try another router...

[az2008]

I forgot to ask. Can someone tell me how to set up QoS on dd-wrt. Since there has been so much discussion about QoS, I'd like to set it up to see if it makes a noticeable difference.

I've used Tomato and DD-WRT. I like Tomato better because it's simpler if all you want is QoS. (DD-WRT has a lot of features if you want to do more things).

However, if your internet speed is variable (slower at night, or your provider uses "turbo boost") then Tomato (and QoS generally) is problematic because you have to set the max outbound (and inbound) speed that the router should use. And, that max has to be conservatively set to the slowest speed you experience on a regular/prolonged basis.

So, if you have DSL (which is famous for consistent speeds) Tomato would work well. But, if you're on an oversold Cable network, and it gets slow at night (and/or, it has "power boost") then you have to set it to the slow speed (not the daytime or "boosted" speed) and never take advantage of the extra speed.

DD-Wrt has a newer Linux kernal with TCP Vegas (enabled by default, I believe. You might ask on the DD-WRT forum.). A lot of people have said this works better than QoS. It's a congestion avoidance algorithm. Also, in DD-WRT you can set the max out/in bandwidth to 0. I think this means DD-WRT will apply QoS to the packets, but not try to cap the overall traffic. (TCP Vegas would supposedly pace the speed to avoid congestion.). I've seen people say it works well without QoS. Others say they use QoS on top of TCP Vegas.

Either way, I tried it and I wasn't too impressed. The quality of the call with Tomato was better. However, I have 1500/250 connection. Maybe DD-WRT would have worked better with a faster connection.

The other thing I didn't like about DD-WRT is that there didn't seem to be a way to set a rule based upon the destination IP address like you can in Tomato. That means you have to set the rules for UDP ports 5060 - 5070, and 10000 - 20000 (which is a lot of ports to give high priority to). DD-WRT's terminology wasn't very clear to me either. I wasn't too sure how the rules should be setup.

EDIT: DD-WRT lets you specify a the protocol RTP. That's supposedly what's happening on ports 10000-20000 when a call is placed or received. But, I've heard using packet inspection like this for QoS can slow the router down.

Another thing I didn't like is that there's no way I could tell to see if the rule is "catching." Tomato has nice graphs showing you how the traffic is falling into classifications. I think the only want to find out with DD-WRT is to reverse the rules so MJ's rules are set to low priority. Then do a speed test while talking to the echo test phone number. I.e., try to force it to be completely unusable and see if it is. If that works, then you know the rule is being applied like you think.

Tomato's QoS is very nice to use. But, if I upgrade to a higher tier with my cable provider (and get turbo boost, etc.), I'm going to have to delve into DD-WRT again to try and get the advantage many people brag about with TCP Vegas.

Mark

[tDogEast]

I too have the 32 second problem for inbound calls. Mine also used to work just fine until about a month ago. I have changed none of the router settings on my LinkSys WRT610N. I have done quite a bit of research and found the following.

Using an application called TCPView from Sysinternals, I find the following MagicJack connection entries. All are UDP entries for open UDP connections:

10.1.1.50:56653 (IP of my local machine)
127.0.0.1:56653
127.0.0.1:61436
127.0.0.1:61437
127.0.0.1:61467
0.0.0.0:61438

Two more UDP entries appear when there is an active conversation, I suppose to support the incoming and outgoing voice channels. The two entries of most interest are the top two for (in this case) UDP Port 56653. This port assignment changes everytime the MagicJack application starts.

Installed Log Viewer from Linksys and turned loggin on in my router. I called into my MagicJack number from my cell. The following is the most telling entry I found as the cause of the "Awaiting Acknowledge" message and subsequent disconneciton:


DROP IN=vlan2 OUT= MAC=00:21:29:c7:xx:xx:xx:13:5f:05:xx:05:xx:xx SRC=67.88.84.6 DST=98.200.xxx.xxx LEN=465 TOS=0x00 PREC=0x20 TTL=48 ID=0 DF PROTO=UDP SPT=5070 DPT=56653 LEN=445

This is an inbound connection to the WAN MAC/IP address of my router. The source is a MJ Server. Notice the source and destination port numbers. The destination UDP is the UDP that I mentioned above that is with the two local machine entries for MJ. I don't know if this is a NAT issue or what, but the other UDP connections work just fine without intervention in the router. This particular entry is being Dropped by the router.

All I have had to do is set up a Single Port Forward for this UDP 56653 to the IP address (10.1.1.50) of the machine that my MJ is on. Everything then works great. Unfortunately, I have to change this rule in the router every time MJ is restarted because the port numbers change.

Just a little more information - when this rule is installed, the message changes somewhat for this connection:

ACCEPT IN=vlan2 OUT=br0 SRC=216.234.79.8 DST=10.1.1.50 LEN=337 TOS=0x00 PREC=0x20 TTL=46 ID=0 DF PROTO=UDP SPT=1026 DPT=56653 LEN=317

The destination port remains 56653 but the source port changes to 1026 from 5070. Also, the destination IP address is no longer the WAN port IP on my router, it is the IP of my local machine. I think all this means is that the router has successfully NATted the connection.

Hopefully someone more network savy can make heads or tails of this better than I. I wish I had a trace like this before problems started so that I could compare.

Anyway - I would really like to solve this. Any Input?

Toby

[az2008]

Anyway - I would really like to solve this. Any Input?

It's really strange because you're describing an attempt by MJ to open a connection to your system. I have no firewall rules at all (either when I was using the original Linksys firmware, or Tomato) and MJ has worked fine since I got it last June.

I wonder if this is specific to the MJ proxy you're connecting to. (Maybe everyone having the problem are in the same geographic area?). You could try pointing to a different MJ proxy. See http://www.dslreports.com/forum/r209511 ... -are-using

I don't see how this can be caused or fixed by the MJ customer. I don't know what you could do to control the server opening a connection to you.

Mark

[tDogEast]

While I certainly agree that the MJ server is trying to open a connection to my PC, it is trying this using a UDP port that the MJ is listening on (in this case 56653). This had to be communicated to the MJ server sometime during the registration when the application first runs for the MJ server to know this port that changes at every restart of the MJ application.

I just don't know enough to understand why the router is blocking this request for connection and why MJ is listening to this port both from my local IP address (10.1.1.50) and the loopback(?) address of 127.0.0.1.

For whatever reason, this "handshake" connection is being blocked and causes the incomming call to timeout after 32 seconds. Allowing this connection to be accepted cures the problem, although manually.

[qhjyz]

I have the same "32 sec" problem from probably a few weeks ago. I have a "WRT54GL" router running very well for more than 5 months.

[az2008]

While I certainly agree that the MJ server is trying to open a connection to my PC, it is trying this using a UDP port that the MJ is listening on (in this case 56653).

I probably don't understand the details of how this works. If the MJ device opened the port and is waiting (listening) to accept a connection, then it requires your firewall to allow that connection.

For example, I believe I can write a script to listen for connections on a certain port. But, if the firewall prevents connections to that port, my script will never receive anything.

So, what you're describing is very different than what I experience with my MJ device. When the softphone starts, it opens a connection to the MJ proxy server's port 5070. That's a control connection. Not much traffic. When a call is placed or received, it opens a connection to the MJ server (not the proxy, but a host with the name "vms1" plus the city.talk4free.com). The port it connects to on that server is in the range of 10000 and 20000. I believe the port it connects to is communicated to my MJ device from the proxy server connection on 5070.

These things may involve my ports in the 50000 range. But, I believe it's all originated from the softphone to the MJ servers as described above. That's what's different from what you describe. You're describing the MJ server connecting to your 50000 port. I'm sure that's not happening on my system because my firewall has no special rules to allow it.

I keep wondering if this is something specific to one MJ server.

Mark

[TrailSeeker]

I am having the same issues with incomming calls being disconnected after 32 seconds and the softphone app says "Awaiting Acknowledgement". I have a linksys wrt610n router and wondering if anyone else got magic jack to work with this router. I brought my magic jack to work and incomming calls work fine so I know its not the proxy server that magic jack uses and that its not my laptop since I am using the same laptop at home and work. Any ideas?

[az2008]

I know its not the proxy server that magic jack uses

I don't think you've proven that. You can prove it conclusively if you follow this, and try the Phoenix proxy (the one I use, and I'm not having the 30-second problem).

http://www.dslreports.com/forum/r209511 ... -are-using

The above write up just explains how to switch to a different proxy1 server. You'll want to follow those instructions, and extrapolate it into "vms1.{city}.talk4free.com" Get the IP address for the vms1 server, and add that to your hostnames file too (just like the proxy1 server, so you fake MJ into going to a different host). The connection to the vms1 server is made when you place/receive a call. It's where the voice traffic occurs.

If you still have the problem, revert to your default proxy.

Then, I'd suggest using wireshark to see if you can figure out what's happening.

It's a strange problem because it doesn't affect a lot of people. People who haven't opened any ports on their routers.

Mark

[az2008]

You can prove it conclusively if you follow this, and try the Phoenix proxy (the one I use, and I'm not having the 30-second problem).

Update. This doesn't seem to work any more. When I wrote that article I could see MJ resolving the proxy1 hostname to an IP address. But, tonight it just uses the IP address. Like it's cached somewhere.

I'll have to examine the registry.

Mark

[TrailSeeker]

az2008 - The magicjack does not work at my house but works at my work. So with that test you are saying that it does not exclude the proxy server at a possible issue? Please explain?

Also, if at home i put my computer in the DMZ, magic jack works. So it must be some port that the router is not letting in. Does anyone know what ports for incoming and outgoing?

Thanks

[az2008]

az2008 - The magicjack does not work at my house but works at my work. So with that test you are saying that it does not exclude the proxy server at a possible issue? Please explain?

This might be a matter of semantics. But, if you use the same proxy server at work, then it doesn't seem like it's the proxy server (if it fails at home, but not at work).

Also, if at home i put my computer in the DMZ, magic jack works. So it must be some port that the router is not letting in. Does anyone know what ports for incoming and outgoing?

It seems like MJ uses different ports for different people. Until a few weeks ago, it used the destination port 5070 (on the MJ server. Maybe 5060-5070) for the control connection when MJ starts. It used a destination port (on the MJ server) between 10000-20000 when a call is placed or received.

Unless I missed something, it seems like a few weeks ago some people began reporting that the MJ server is trying to open a connection to their computer.

I don't think this will be understood until someone affected takes down their firewall and analyzes what's happening using wireshark.

It's strange because one person said that flashing his router with Tomato fixed it. That indicates it's not the MJ server trying to open a port.

Mark

[TrailSeeker]

Thanks for the quick reply. so to verify the ports and direction:

Port Range Type Source Destination
------------- -------- ---------- ---------------
5060-5070 udp my computer MJ server

10000-20000 tcp MJ Server my computer


is the above chart correct?

[az2008]

Thanks for the quick reply. so to verify the ports and direction:

Port Range Type Source Destination
------------- -------- ---------- ---------------
5060-5070 udp my computer MJ server

10000-20000 tcp MJ Server my computer


is the above chart correct?

What I see is that both are udp, and both are to the MJ server.

It appeared to me the MJ proxy1 server (control port on 5070, or maybe 5060-5070, but I only saw 5070) passed the server name or IP address of the vms1 server to my computer and it opened a connection to 10000-200000 on the MJ vms1 server.

That's why I have no ports open on my router and MJ works fine. It's why I never understood dead chat's instructions to customers that they open ports on their routers. And, why I don't understand recent reports (like yours) that you can't get MJ to work without putting your computer outside the firewall.

Mark

[TrailSeeker]

It could be a bug in my router as well... I use to have an airlink 101 router and did not have this particular issue. But I had to replace it due to bad wireless signal. It also looks like linksys might have a bug maybe for voip applications? Not sure.. but I can tell you this is one issue that is annoying. I do have an option for port triggers in my router.. now my understanding of this is that you define outbound and inbound ports and the router track which ip address requests what external ip on the web and allow that external ip to open a port back to the internal ip since it initiated the request. Now with all that said.. I thought routers did this by default where it knows the ip I request and allows that external ip to open a port back to me since I was the one who initiated it.

[TrailSeeker]

What I see is that both are udp, and both are to the MJ server.

So both ranges are my computer opening ports to MJ servers? and MJ does not open a port to send back an acknowledgement?

I thought this is like ftp.. your outgoing port is 21 but the ftp server on the internet will open a random port in a particular range to send back acknowledgements.....

[az2008]

I thought this is like ftp.. your outgoing port is 21 but the ftp server on the internet will open a random port in a particular range to send back acknowledgements.....

From what I've seen with my computer, it's more like FTP after you issue the "passive" subcommand. That subcommand was added to FTP precisely because of these firewall problems. Port 21 is the control/command channel. When you do anything that will generate data (like a dir or put command), the server tells your client (on the control port) the port your client should open on the server for the data. The server then begins listening on that port, waiting for your client to connect.

Regarding your router being buggy. I don't know. Most people with this problem seem to be using a WRT54GL. That's what I use, but mine works without opening any ports for external connections. It could be a bad batch if everyone just bought new routers. But, one guy said his WRT54GL worked fine for months, and just began having this 30-second problem a couple weeks ago.

That makes me think the common thing among everyone experiencing this is the server(s) they connect to. There is a way to change the server you connect to. (Wipeout97 just posted that you can specify IP addresses in your hosts file. It will be awhile before I can test this and update my write-up to include it. If you can put it together for yourself, you could try it. http://www.phoneservicesupport.com/post24003.html#24003).

Someone else (maybe the same guy who said his router worked fine for months) said flashing their router with Tomato fixed it. If you have a WRT54GL, you could do that. It's a better firmware anyway. You'll be able to use QoS to improve the quality of your calls. Just be careful. You can brick your router. You'll want to be sure to read all the instructions and warnings. DD-WRT has many more warnings, such as only using IE to do the flash. I tend to keep some of those things in mind and follow them when I flash Tomato (including resetting factory defaults first, clearing NVRAM after the flash, maybe before too.).

Mark

[qhjyz]

After I install Tomato firmware in my Wrt54GL, my MJ works fine now.

[lschutz]

I have not been able to receive incoming calls consistently for the past few months. I tried MJ Chat and was on hold for nearly 2 hours before being cut off. I tried three different phones and all the MJ FAQ and help suggestions to no avail. Curiously, the MJ works fine when it is rebooted through the Advanced Users part of the Menu, but only works for a short while. After reading a number of blogs, I tried disconnecting the router and directly connecting the ethernet cable to the computer. So far, it seems to be working fine. I have received several calls with no problem and no incoming calls have been unsuccessful. I have my fingers crossed that this will continue to work. It's not much of a fix if you need your router for wireless, but we are hardwired throughout the house, so we can do without it most of the time.

[toddnoch]

I want to thank the host of this site and particularly everyone that has participated in this discussion. As my subject reads, I was having two major issues with my MJ.

The first was on my Windows Vista 64-bit computer. It was dropping calls at exactly 5:12 into outbound calls. On my Windows XP 32-bit computer, it was dropping them at 15:13 (or somewhere around there). That issue turned out to be my Avast! free home use anti-virus software. Once I went to "Program Settings...", I clicked on "Exclusions" and added the two MJ fixed drive letters as well as "C:\Users\{username}\AppData\Local\tjnet" & "C:\Users\{username}\AppData\Roaming\mjusbsp" folders. You'll need the appropriate paths translated on an XP computer.

The second issue is what's being discussed here. I thought I was in good shape until my wife pointed out that it woundn't last more than 32 seconds on an inbound call. I tested it over and over again, and ure enough she was right (she always is)! Well, I disconnected my router and went right into my DSL modem and all was well. That confirmed IT WAS MY ROUTER! I have the Linksys WRT54GS. I was initially hesitant to flash the bios, but it couldn't have been easier. Just know your previous settings and most of all your ISP login credentials. Also, please heed all the warnings of potentially "bricking" your device as it could become unusable if done incorrectly! I flashed mine with Tomato (and plan on donating to the cause) and love it. It fixed the problem right away. I believe the issue is caused by a "bug" in the Linksys firmware. Remove the firmware, remove the problem.

I also recently purchased the Hawking Broadband Boaster, which is a Quality of Service (QOS) device in a box. Although you can alter the priority of data (i.e. what gets through first, VOIP, games, torrents, web surfing, etc.), this device is made so you won't have to. It's basically plug-and-play. I know the Tomato firmware has this feature as well, but I didn't want to have to figure it out. Now, I can be downloading at 600kb/s, upload at 40kb/s then surface the web and place a call without incident! The Hawking BB will throttle the "less important" requests down and allow the higher demand items such as VOIP, video streaming, and web surfing to take priority over my less important torrents.

I hope my experience helps others as the previous posts really helped me. And, for all the people who said MJ was a scam and didn't work - the best thread I read somewhere was someone pointing out that it works just fine. If there's an issue, it's likely being caused by something else in your system. I stuck to it and exercised the "art of troubleshooting" and resolved it on my own. Unfortunately, the MJ chat support really was that helpful. They didn't even suggest any of these items.

[GreyGhost]

Had the same problem with the Pending ACK. I went ahead and flashed the WT610N with dd-wrt.v24-11762_mega-wrt610n.bin. Everything is working great. I even think the performance of the router is better but need to get tools that prove this. I did use the newest build of the dd-wrt because I found that this build had work for several people. When I feel brave I'll try the latest build.

LINKSYS Firmare on 610n is a problem for SIP

[tDogEast]

Linksys has finally posted new firmware for the WRT610N. This fixes the MagicJack Pending ACK problem for this router.

The new firmware lets you turn OFF (Disable) SIP ALG. The SIP application layer gateway was included in the original code, but since MJ implemented it's own methodology for dealing with SIP behind a NAT router, the ALG in the router did nothing but interfere.

Simply go to the Administration tab in the router's web interface under advance features and DISABLE SIP ALG.

After I did this, no more problems. I also no longer have to port forward 5060 and 5070. Everything is working as it should for now (only took Cisco 6 months to fix!).

[azoo]

Is there a new and easy way of fixing this ACK issue. I have read all the flashing tomato , new firmwire etc, but I am not good with computers and I know my router is a linksys and i am having this pending ack issue. can someone help me with an easy fix ? My MJ worked fine for the last 2 weeks but started this ACK issue a few days ago. is there any other way of fixing this issue ? without touching the linksys router ? please help.