Vera Edge has no internet access

Thanks

I rebooted Vera and ran the date command in terminal:

Sat Oct 10 23:38:49 BST 2015

Looks to be the correct date and time.

I then ran the /etc/init.d/dnsmasq restart command a few times

Looking in Net & WIFI it still say No Internet Access

EDIT: Although I am not getting the spinning wheel of death at the moment. But still no internet access for Vera or DNS working.

Just to clarify:

/tmp/resolv.conf.auto has your router’s IP address in it. You restart dnsmasq and you still don’t have internet access?

I don’t know what’s up.

Yes that’s exactly correct

/tmp/resolv.conf.auto always seems to have the router’s IP address in it.

And restarting dnsmasq does not bring back the internet access.

Cheers

Good detective work to identify that it’s a DNS problem.

Another command that you can add to your arsenal is nslookup. It’s a command-line DNS resolver.

nslookup www.google.com 192.168.0.1

The last term is optional; nslookup falls back to whatever is in /etc/resolv.conf (so localhost on the Vera). Try leaving it out; try putting your router’s IP address there; try putting a public DNS server like 8.8.8.8 there.

I’m having a similar problem that cropped up on a vera lite that has been running untouched for over a year. Yesterday morning it just lost internet connectivity for no apparent reason.

I ssh in and can ping out to say 8.8.8.8 but nslookup fails. Clearly the DNS is jacked up. I tried comcast dns and google DNS both with no joy. I did try restarting dnsmasq last night but that didn’t do anything either. I haven’ checked the /tmp/resolv.auto as I didn’t know it existed. I looked at /etc/resolv.conf on a working vera and the broken one and note they are the same so I assumed it wasn’t using that file and moved on.

I’m running UI5 on it.

I opened a ticket with Vera and we’ll see where it goes.

Think support have fixed mine, I’ve not tested much as been busy but after rebooting the Edge unit once the internet access was still working on it ok.

Here is what they said was the problem.

The reason why the /etc/resolv.conf file is rewritten on each reboot is because there?s a symlink to the temporary folder on the unit. With each reboot the file is recreated in the /tmp/ folder of the unit with the default values, and your changes will be deleted. You can see the symlink below :
lrwxrwxrwx 1 root root 16 Jun 23 13:08 /etc/resolv.conf → /tmp/resolv.conf

To fix this you?ll need to break the symlink by rewriting the file and I?ve done that for you.

@cw-kid - Please test and let us know. On working Veras /etc/resolv.conf → /tmp/resolv.conf only contains

search lan nameserver 127.0.0.1

This makes Vera query its local dnsmasq for name resolution. But, at least when Vera is using DHCP, the dnsmaq process looks to /tmp/resolv.conf.auto for its name server information. This is specified in the startup command with -r /tmp/resolv.conf.auto which can be observed by looking at the process list with ps ax. In other words, under normal circumstances, DNS resolver infromation from your DHCP server(router) should be found in /etc/resolv.conf.auto and no where else.

It sounds to me like support fixed your ability to save changes to /etc/resolv.conf. This would be an effective workaround to your issue, but it doesn’t fix the original issue that you described.

I thought the same thing but was waiting to look at my misbehaving unit first.

I’m wondering if my .auto file is jacked or something.

So you want me to confirm that /etc/resolv.conf and /tmp/resolv.conf only contains 127.0.0.1 ?

[quote=“cw-kid, post:29, topic:189162”]So you want me to confirm that /etc/resolv.conf and /tmp/resolv.conf only contains 127.0.0.1 ?[/quote]Yes. I’m trying to determine if your Vera is working as a default Vera would, or if you and support have created a workaround to the problem by forcing the bypass of dnsmasq.

[quote=“Moxified, post:28, topic:189162”]I thought the same thing but was waiting to look at my misbehaving unit first.

I’m wondering if my .auto file is jacked or something.[/quote]It may be, or it may be something else. At an earlier point in this thread, @cw-kid had the files set correctly, but dnsmasq still was not resolving addresses, for reasons we have yet to determine.

The etc/resolv.conf file has the routers IP in it

search lan
nameserver 192.168.0.1

The tmp/resolv.conf file has

search lan
nameserver 127.0.0.1

And the tmp/resolv.conf.auto has

Interface wan

nameserver 192.168.0.1
search home

I have just pulled out the power lead from the Edge, first time I have done this since support were connected to it. After it came back up and I could access the web GUI, I looked in Settings - Net & WIFI and it still says Internet Access is OK.

Looking in the those three files again:

The etc/resolv.conf file has

search lan
nameserver 192.168.0.1

The tmp/resolv.conf has

search lan
nameserver 127.0.0.1

tmp/resolv.conf.auto has

Interface wan

nameserver 192.168.0.1
search home

So they are the same as before the power cycle.

So is that good or bad ?

For you, its going to work. But, it’s not the standard behavior. It probably won’t be an issue for you unless you change your LAN’s subnet to something other than 192.168.0.x At that point you’ll have to manually edit /etc/resolv.conf

In my case, I checked and the resolv.conf and .auto were correct. I replaced nameserver 127.0.0.1 with nameserver 8.8.8.8 and it started working again. Obviously on reboot it broke. Clearly dnsmasq is broken. I stopped dnsmasq and tried starting it again but it complained something was already using port 53.

This is at my father’s house so I didn’t want to take too much of his time over webex. I just left it with 8.8.8.8 so that he would have remote connectivity to his system again until I can get support to look at it. I can also initiate a support session to it now.

I’m wondering if I will have to revert to factory and restore a backup to get the dns service binaries and configurations all fixed.