01-20-2014 10:38 PM
Hawaiian Telcom just "upgraded" my Pace 3081 RG with a Pace 5168NV RG and now I can't keep a call going on the microcell, AT&T sent me a new microcell, thinking the problem was the microcell.
The problem is any call, in or out bound, at about 40 seconds, the 3g light turns off, and the call fails. My iPhone 5 shows a call failed message, and the 3g light starts flashing and will be solid after about 1 minute.
From the Trouble shooting guide, I have set the MTU to 1492.
5 days of calls to AT&T microcell support have been uselss, they have no idea what needs to be changed on the RG. They say to call Hawaiian Telcom, and Hawaiian Telcom says to call AT&T.
My internet is 23/5.
01-23-2014 10:38 AM
To be honest, I never bothered with MTU settings so I don't actually know what mine is set at. It just always worked. Isn't 1492 a default for routers? Maybe that's why mine always worked. Yeah, I'd have the OP look into the other required settings. We sometimes concentrate on the ports and forget about IPSec Pass-through, MTU, and Block Fragmented Packets.
01-23-2014 10:47 AM
The default MTU for HT is 1500, on the 5168, I can't find any place to configure either the IPSec Pass-Through or Block Fragmented Packets.
Last night, HT had me make a call over the Microcell whiile they did a packet capture, call failed like normal. The CSR did tell me they have a ticket open with AT&T.
01-23-2014 10:51 AM
Well that's something. At least they're trying. So basically, your MicroCel was fine until HT upgraded their network and gave you a different modem, right? MTU/packets/IPSec are all important so maybe with their new modems/upgrades they just need to tweak their end a bit. Hopefully AT&T can help/advise them with that.
01-23-2014 10:55 AM
A MTU of 1492 is the standard setting for Ethernet with LLC (Logical Link Control), SNAP (Subnet Access Protocol) and PPPoE (Point to Point Protocol over Ethernet). In other words, a very common maximum packet size setting. So, 1492 bytes is the maximum packet size that can be sent without fragmentation.
If the ISP is using a MTU setting that is smaller than that (and they could) and the Mcell network needs a MTU of 1492 (which I don't know is a fact) and fragmented packets are being created as a result, then they could be blocked by the gateway if blocking fragmented packets is enabled.
01-23-2014 11:00 AM
If HT is using a MTU of 1500 then I'm out of ideas except fragmented packets. I hope the OP can continue to get cooperation between HT and AT&T to solve the problem.
01-28-2014 12:45 PM
I got a call from AT&T yesterday, the person was calling to see if the problem was resolved.
It seems a bit odd that, from what the AT&T person told me, they really don't have any way to find out what happens when the Microcell disconnects, or any other issue. Their tools just don't have any info.
Hawaiian Telcom has told me that Pace is working on the issue. HT did a packet trace on their end while I tried a call through the Microcell. It took a little of 2 minutes to fail, but it did.
The 2nd tier CSR at HT want to help, he want's to get a Microcell for home, but now is holdng off until this is resolved.
There is a cool tool, pathtrace, for Linux. It traces path to destination discovering MTU along the path.
It seems that no mater what IP or domain name, the PMTU is always 1500.
01-28-2014 12:47 PM
This morning, I called Hawaiian Telcom for any update, and was told Pace is sending a technical representative to Honolulu to work on the issue. It seems that there are lots of microcells on Hawaiian Telcom's network that have all stoppped working.
Pace firmware on the 5168 is 10.1.2.494555-ht, currently.
01-28-2014 2:26 PM
Thanks for reporting back. It might be as simple as an MTU issue and the inability of the Pace modems to be adjusted down to 1492. I'll pass this along to the Admins and see if they can followup as well.
01-28-2014 3:09 PM
One of our Managers will be reaching out to you within 48 hours at the email address listed in your profile here on the forums. If there is a better way for us to get in touch with you, please reply via private message with your contact information.
AT&T Social Media Manager
02-05-2014 12:50 PM
Ok, so, Hawaiian Telcom has verified that this is a Pace 5168NV firmware issue.
There is also an NTP exploit in the firmware, and Hawaiian Telcom is blocking all NTP servers execpt the ones that the Microcell is attempting to reach.
Thanks to wireshark and a network tap, I was able to find out what the NTP servers IP's were.
The Microcell was reporting "The registered address does not match the actual device location." FTC231 or 233, but it was really the NTP servers could not be reached.
As soon as Hawaiian Telcom opened the 4 Ip's (18.104.22.168, 22.214.171.124, 126.96.36.199 and 188.8.131.52), the Microcell "actviated". AT&T had removed and attepted to re-active several times until I found the NTP issue.
DMZ Plus mode does not even work, that's how broken the current firmware is.
Pace is currently working on a new firmware version and I should hear back soon that they want to push it to my RG.
02-05-2014 12:58 PM
Thanks for the update. Firmware issues have been an ongoing issue with some of the Pace modems so I'm glad that they are taking an active role in resolving your issue.
- edited 02-06-2014 9:35 AM
I'd take an active role also if I were a Pace tech and had to go to Hawaii to resolve a problem!
02-12-2014 12:01 PM
Some small updates.
Pace did not send anyone to Hawaii :-(
Hawaiian Telcom swapped out the 5168 for a 3801 and of course it works.
They asked if I would keep the 5168 as they would like to use me a a test person for the new firmware that Pace delivers. And they just updated the provisioning to 2SD/4Hd Wan but still 2SD/2HD Ingress, but that only works with the 5168 here.
Update today, Hawaiian Telcom is now looking to drop the Pace 5168 and move to yet a newer version. They did not give me the model number, but supposedly the new Pace works with the 3G Microcell and has the NTP bug fixed. HT is trying to obtain a unit for me to test with. Lets see how long that takes.