Mifi: LTE connected but no Internet

All,

I have a serious issue with the Mifi. I bought one with an internal battery and the EC25 module for Europe.
First thing I did was to update the firmware and upgrade it to the official lede-2.271. I then did the basic configuration and connected via wifi. Form there I wanted to set up the 4G connection. I put in my Telekom SIM which I tested before in a mobile phone. LTE and everything was working. I removed the PIN from it and configured the “modem” interface. It detected the modem model. the IMEI of my card and the carrier. I checked the automatic connect and after a few seconds I got a “connected” and the IP address in the webfrontend, but no Internet at all.

I logged in via SSH and checked with uqmi -d /dev/cdc-wdm0 --get-current-settings and got this status:

{
“pdp-type”: “ipv4”,
“ip-family”: “ipv4”,
“mtu”: 1500,
“ipv4”: {
“ip”: “10.29.28.1”,
“gateway”: “10.29.28.2”,
“subnet”: “255.255.255.252”
},
“ipv6”: {
},
“domain-names”: {
}
}

The route is also correct:
root@GL-MIFI:/# ip route default via 10.29.28.2 dev wwan0 src 10.29.28.1 text

The only thing that is missing here are the DNS servers. So I thought that would be the problem, but it is not. I am able to ping the default gateway (in that case 10.29.28.2), but not something like 8.8.8.8.
I tried almost everything from re-configuring and re-connecting but nothing helped so far. I also disabled the firewall, added DNS servers myself, changed the network config to everything possible, but no luck.
Below my network config in /etc/config/network. The modem config below was created via the automatic setup, but I also tried everything with username, password, auth-type or just proto ‘dhcp’ and otion ifname ‘wwan0’ etc.

config interface ‘modem’
option ifname ‘wwan0’
option service ‘fdd_lte’
option apn ‘internet.t-mobile’
option proto ‘qmi’
option device ‘/dev/cdc-wdm0’
option metric ‘40’
option disabled ‘0’

Does someone know what I am missing here? I also talked to Telekom and they say they don’t block anything and that there is everything enabled for that SIM card.

silly question: what if you just turn off then turn on the mifi? Will it work?

Unfortunately not. I also tried that several time thinking that the SIM card may not be recognized correctly etc.
So I powered it completely off, powered it on again and same issue. Maybe there is a super special setting for connecting to Telekom in Germany? I mean like in one of the other threads with AT&T in US, but here there was no IP assigned at all and when I talked to Telekom they said the settings are fine. Mifi with EC25-A AT&T USA connection - #7 by edhorton_01

Really lost at the moment :frowning:

can you ssh to the router and check if the router can use the Internet.

ssh root@192.168.8.1

$ ping 8.8.8.8

No that is not possible. That is what I also posted in my first post:

I am able to ping the default gateway (in that case 10.29.28.2), but not something like 8.8.8.8.

that is a problem. I have no idea.

Did you have any other cables connecting to the WAN port?

No. The only thing that is currently connected is a UART to USB adapter in order to log in via serial and of course the micro usb cable for power.

can you try again adding wwan0 as interface

ping -I wwan0 8.8.8.8

Does also not work :frowning:

root@GL-MIFI:/# ping -I wwan0 8.8.8.8
PING 8.8.8.8 (8.8.8.8): 56 data bytes

@rene, any chance you can have a check? Seems everything is fine.

Just want to mention that I ordered a second one which should arrive on Monday. Maybe the first one has some hardware failure if you say that all settings are correct.

So the new one I received two days ago and it was working out of the box, BUT with the 3g module instead of using the qmi mode and therefore using /dev/ttyUSB2 instead of /dev/cdc-wdm0. So I reconfigured the other one to the 3g mode and also using /dev/ttyUSB2 and it was working as well.

Then I switched the new one in the webinterface to 4g, then using qmi and /dev/cdc-wdm0 and it was still working. So I did the same to the other one and this is now working as well with that setup. I didn’t change anything else so could there be the possibility that something on the EC25 chip changed or is set/removed while switching from the 3g-module to qmi? Or is there a reset command for the chip that can be issued in qmi mode that could have caused this issue?

I only want to get to the root cause of this so that no one else is stuck at the same point :slight_smile:

Config set (out-of-box)

config interface ‘modem’
option proto ‘3g’
option ifname ‘3g-modem’
option device ‘/dev/ttyUSB2’
option country ‘country21’
option sp ‘isp2’
option dialnum ‘*99#’
option apn ‘internet.telekom’
option username ‘telekom’
option password ‘tm’
option metric ‘40’

Config now set in qmi mode

config interface ‘modem’
option ifname ‘wwan0’
option service ‘fdd-lte’
option proto ‘qmi’
option device ‘/dev/cdc-wdm0’

Seems this is a bug in qmi script and didn’t get fixed. The script affect the modem. Seems what have done get the modem fixed.

Okay thanks. That makes it clearer. And does that mean it has to be fixed on your end?

Yes. We need to fix.

Hi! Was this issue fixed? I’m asking because I’m trying to follow the poster’s instructions on moving back to the 3g module, as it appears I’ve screwed something up using uqmi commands. I’m unable to restore a connection to my lte provider now on /dev/ttyUSB2. Here’s what’s showing in my log file (looks like lots of duplicate ‘modem’ iface entries, but only one when I check /etc/config/network).

Mon Jul 9 23:34:09 2018 daemon.info gl-health>>: AP Ok=1
Mon Jul 9 23:34:09 2018 daemon.info gl-health>>: network(caIfnames) = wan wan6 tethering modem modem modem modem modem modem modem modem modem modem modem
Mon Jul 9 23:34:09 2018 daemon.info gl-health>>: STA Ok=1
Mon Jul 9 23:34:09 2018 daemon.info gl-health>>: Everthing is OK!!!
Mon Jul 9 23:34:09 2018 daemon.info gl-health>>: check wpa runing =
Mon Jul 9 23:34:09 2018 daemon.info pppd[19220]: Exit.
Mon Jul 9 23:34:09 2018 daemon.notice netifd: Interface ‘modem’ is now down
Mon Jul 9 23:34:09 2018 daemon.notice netifd: Interface ‘modem’ is setting up now
Mon Jul 9 23:34:11 2018 daemon.notice pppd[19343]: pppd 2.4.7 started by root, uid 0
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: abort on (BUSY)
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: abort on (NO CARRIER)
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: abort on (ERROR)
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: report (CONNECT)
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: timeout set to 10 seconds
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: send (AT&F^M)
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: expect (OK)
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: AT&F^M^M
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: OK
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: – got it
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: send (ATE1^M)
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: expect (OK)
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: ^M
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: ATE1^M^M
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: OK
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: – got it
Mon Jul 9 23:34:12 2018 local2.info chat[19419]: send (AT+CGDCONT=1,“IP”,“fast.t-mobile.com”^M)
Mon Jul 9 23:34:13 2018 local2.info chat[19419]: timeout set to 30 seconds
Mon Jul 9 23:34:13 2018 local2.info chat[19419]: expect (OK)
Mon Jul 9 23:34:13 2018 local2.info chat[19419]: ^M
Mon Jul 9 23:34:13 2018 local2.info chat[19419]: AT+CGDCONT=1,“IP”,“fast.t-mobile.com”^M^M
Mon Jul 9 23:34:13 2018 local2.info chat[19419]: ERROR
Mon Jul 9 23:34:13 2018 local2.info chat[19419]: – failed
Mon Jul 9 23:34:13 2018 local2.info chat[19419]: Failed (ERROR)
Mon Jul 9 23:34:13 2018 daemon.err pppd[19343]: Connect script failed

Any ideas?

Same problem as above, any idea on how to resolve?

I am not sure if the modem has problems.

Anyway you can try v3 firmware for MiFi which is totally new. Not sure if the QMI problem was fixed. We fixed several bugs in QMI scripts already.

Thanks for this, I have a MiFi device that was working fine, then it suddently stopped with this exact same issue. 4G connected, I couldn’t ping anything, but curiously nslookup worked using 127.0.0.1 or using one of the dns servers shown in the output from:

uqmi -d /dev/cdc-wdm0 --get-current-settings

so clearly something was getting through. No other DNS servers worked, however.

Anyway, the fix was simply to open the web interface, manual config of the 4G device, change the device from /dev/cdc-wm0 to /dev/ttyUSB2, and then it immediately worked. After that I changed back to /dev/cdc-wm0, and it’s still working.

Thanks for sharing!

Seems the modem has some problems and it fixed itself when you set up manually.