Port forwarding drops out every night

[#1]

I have a Sagecom 2704n with FTTC and every night between 03:00 and 03;15 all external access stops working. This is true of port forwarding with default firewall and also when setting DMZ on with the firewall disabled. I am running a port check every 15 minutes from an external server and use other port check tools.

I have WiFi and DHCP disabled on the 2704n and the static LAN address is in the 10.x.x.x range. I use a permantly on Synology NAS to provide DHCP and the HTTPS and VPN services I am trying to access. I also have a couple of bridge mode routers with WiFi enabled in more suitable locations than the 2704n.

Rebooting the 2704N restores service until the next night but is useless if I am away from home which is when I expect to use remote access.

None of the posts here mention what seems to be a scheduled failure of the router. Has anyone come across this and solved it?

[#2]

I have no experience of the specific router but this is odd. The only thing that occurs to me is that around 3 am is often the time daily scheduled tasks run. Also, rebooting the router will also reset all connections routed through it, which could affect other devices, so the blockage might not be in the router itself.

You do not give sufficient information about how your system is laid out to enable me to think about it further, nor of the OSes on the devices connected to your router, so people are unlikely to be able to suggest anything intelligent. All we know is that the forwarded ports stop working around 3 am and that rebooting the router seems to enable them again.

[#3]

I was hoping that others had experienced similar regular drop-outs and had a solution. More detail:

The Sagecom 2704n router was provided by Plusnet, to replace a Technicolor router that they had installed (at the same time as a fibre modem), to support the multicast feature required by YouView TV.

I have modified the settings of the router to set a different LAN subnet rather than reconfigure my intranet, and have disabled DHCP as the configuration options did not permit me to override the DNS server address. Wifi is disabled. The 2704n is connected directly to a Linksys EA6700 router that runs in bridge mode and provides Wifi and onward switching not routing. The 2704N also is connected to a Linksys SPA-3000 VOIP device.

There are various Raspberry Pi devices (Debain based) , a couple or W10 computers, a Sony TV, a YouView box, some Wifi phones/tablets, occasional laptop(s) and a Synology NAS (arm processor, Debian based OS - DSM6) connected to the EA6700 via further switches as required.

The network works OK and access to the outside world, including access to Dropbox, hubiC and Amazon S3 from the Synology NAS are not affected by the drop out of external inbound access at just afer 03:00.

A second Synology device that starts up at 02:45 and is used to back-up the primary Synology NAS at around 3 a.m. has been discounted by running tests at different times, however to be totally sure I have changed the schedule, from today, to avoid the 03:00 time.

After the drop out, the ports are always still accessible from local devices on the intranet, so it points the finger very much to the 2704n.

Pete

[#4]

Am I right in thinking the Synology NAS is the destination for the port forwarding? If so, has it any packet-filtering/shorewall or other firewall of its own?

If cron.daily runs on the NAS would it be worth changing the time that runs to see whether that changes the time the ports drop out?

[#5]

Yes, the ports are forwardrd to the Synology NAS. It does have a firewall, but that is configured to have the ports open to all addresses. Internal access to the ports is never affected anmd no action, apart from restarting the router, is required to enable external access.

I have just found that deleting and resetting up the HTTPS fort forwarding rule on the 2704n re-allows the external access when it is blocked, so it really looks like it the port forwarding that is being affected every night,

I haven't been able to find cron daily in the NAS - the root's cron has some stuff in it but the only thing about that time is a ntpdate. Running that manually doesn't affect the forwarding, but I have changed the time anyway to definitely eliminate it. There are no other crontabs in /var/spool/cron/crontabs

Tomorrow night's test is to isolate the 2704n from the rest of the internal network overnight, reconnect in the morning and check the port forwarding - I won't know when it stops, if it does, but it would remove suspicion from all internal devices.

Pete

[#6]

That sounds like a good first test. However, if deleting and remaking the rule also restarts the forwarding that itself lets other devices off the hook.

In terms of troubleshooting, I don't know whether there's anything interesting at https://community.plus.net/t5/Tech-Help-Software-Hardware-etc/Unlocking-the-potential-of-Sagemcom-27.... Ideally you need to look at the logs to see what the router thinks it's doing at the time the forwarding stops.

[#7]

That thread was very interesting. At the moment I am a bit of a limbo - there were a number of port forwarding drop outs during the day and evening on the 24th and 25th followed by a, so far, final overnight drop out at 03:00 on the 26th. This was after I had ensured that no scheduled tasks were set to run at that time. I am continuing to monitor the ports to see if it starts happening again. Not a satisafactory conclusion at this time. Thanks for your help - I may be back.