Fibre connected Sky Hub SR203 randomly rebooting

[GilesGodartBrown]

Since the middle of September the Sky Hub (SR203) has been rebooting randomly between every few seconds to a number of hours, but many times a day (9 times in last 24 hours)

I know its rebooting from its log which shows it starting with the 1700 time, until it contacts NTP. 

I know when it is happening because I have a process pinging it every 5 seconds. and Raspberry Pi watchdogs  also detect it

When it reboots I lose the Broadband and telephony connection for about 90 seconds, it then comes up for a few seconds then drops again for about 10 seconds.

I live in a remote area with very poor mobile coverage, so voip and Mobile Wifi calling are essential in case of an emergency.

Here are the steps I've taken following several calls with Sky, none of which have had any effect.

Bypassed the BT Openreach power pack which was showing an error. No errors on the ONT itself.

Factory reset the Sky Hub

Swapped the Sky Hub for a replacement from Sky

Checked the AC power - the Hub is on a UPS and other devices on same UPS show the power is good

Disconnected devices - Amazon Echo, cctv, Sky Mini

Turned off Wi-Fi on the Sky Hub, Sky Q and Sky Minis (they are all wired via a GiGE switch).

Run the Broadband checker in the My Sky App - this finds no errors.

Checked the public IP address - its getting some really odd public adresses, sometimes on the 2.22x.xxx.xxx network, but changes to the 90.202, the new router connects to the 130.6xx.xxx.xxx and 132.xxx.xxx.xxx networks.

Anyone got any further suggestions?

[cookiemonsteruk]

@GilesGodartBrown 

Could you post your router statistics

https://helpforum.sky.com/t5/Broadband/How-to-find-your-Sky-Broadband-router-stats/m-p/2855717#M1014...

[Chrisee]

@GilesGodartBrown are you sure the hubs are rebooting rather than losing connection and reconnecting the stats wil show thst the up time resets each time the unit boots? A full reboot of a SR203 takes around 3 minutes and can be monitored by watching the LEDs. see  https://www.sky.com/help/articles/hub-lights-explained-start. If the unit is rebooting spotaenously that is a fairly common indication thst it needs replacing, To have two units showing the same fault means it is over whelmingly likely that the reboots are not due to the hub itself. This leads me to think the connection is dropping perhaps due to an issue with the ONT. Certainly  you should not have bypassed its power supply it is Openreach's property if the power supply is faulty they will replace it.

The range of public IP4 addresses is simply due to the fact Sky have had to buy numerous small address blocks as their customer numbers have grown due to the fact that the supply of IP4 addresses is exhausted so addresses are traded. The internet should be using IP6 by now but many sites and services seemed not to get the memo.

[GilesGodartBrown]

My ONT had a separate Power pack (I assume some sort of short term UPS) Sky advised that BT no longer support this and that I should plug the ONT directly into the PSU used for the power pack.

The Router statistics uptime and observing the lights (all off except power) matches when I believe the router is rebooting.

Stats below;

System Up Time: 03:25:37

Port Status TxPkts RxPkts Collision Pkts Tx b/s Rx b/s Up Time

WAN WANoE 393518 377770 0 54533 22644 03:22:43

LAN Up 488915 551699 13 468946 78342 03:25:37

WLAN (2.4 GHz) Up 895 0 0 2597 0 00:02:49

WLAN (5 GHz) Up 1253 4294967295 0 6065 0 00:02:30

Yes, I've only just re-enabled the WiFi.

I think it may be due to a lot of devices using MDNS so I have now disabled this, and have crossed my fingers.

[GilesGodartBrown · Accepted solution]

I think I've solved it.

It appears the Sky Hub is intolerant of MDNS messages.

I've disabled MDNS from all of my devices, replacing it with hardocded IP addresses and the Sky Hub has been up for over a day now; the longest its been for a very long while.

[BrianMul]

HI There, Giles, thanks for your advice on this. I am having a similar problem with random reboots.  Would love to know if the mDNS issue prooved to be a long term fix.  I got a second SR203 from Sky for my 1Gb SIRO FTTH and it is also rebooting.

I noticed it may be worse when my son is also working from home, perhaps his WIndows laptop is sending a lot of mDNS queries.

I thought about creating a rule to block port 5353, but need to research it more

[BrianMul]

I have created a service mDNS in Security->Services->Add Service 

Then I added a firewall rule to block all mDNS

[GilesGodartBrown]

I stopped all of my IoT devices from using MDNS and the time between reboots has dropped significantly to just about twice a day. 

I'm convinced that the Sky router isn't really up to a large smart home system (>90 Wifi devices), not least of which in that it seems to stop traffic between the wifi side of the network and any wired devices (like a smart home server running Home Assistant for example) when SkyQ devices are acting as repeaters. 

Consequently I have wifi turned off on all Sky devices (hub, Q, mini) and run separate wifi gateways connected to a GiGe switch for my IoT devices.

Annoyingly every now and then the SkyQ devices turn back on their WiFi's, I assume after software updates.

The Sky Hub also reboots regularly if its near another WiFi repeater.

I think the ultimate solution, if I have to keep the Sky router, is to place another router between it and my home network. ☹️

[TheBlackSheep]

I'm getting the same random rebooting on this router - it seems to panic under any kind of stress and as I work from home it's affecti g my work.  Additionally, mM two SkyQ (not on Wifi) remote boxes deactivate themselves randomly (one maybe once twice a  week and the other probably once a day).  Most of the time they require unplugging and rebooting as the reactication menu option can't be selected.  It's all very, very frustrating.

I have quite a few TAPO devices (plugs, lights, cameras) and a few IOT/RasPi devices (I don't run HA).  Most of those aren't connected by Wifi to the router (I have two TP-Link Access Points) but they do get their IP address from the Sky SR203 router.

I'm planning to move to static IP addresses where I can and stop running DHCP on the router other than for about 30 devices if the limits on the router are so low. 

Would it be better to turn off DHCP on the router completely and let some other device provide IP addresses?  I have several Linux boxes that are probably better suited for that anyway.

I did try the mDNS firewall suggestion from BrianMul but then one of the remote SkyQ boxes couldn't connect to the main SkyQ box so I panicked and removed the setting only to find the main SkyQ box had stopped responding and needed rebooting.

[GilesGodartBrown]

My Sky router reboots at least once per day (twice already today) I, like you, am pretty certain its a traffic issue where the router 'just gives up', reboots and you are without broadband for up to a minute, and in my case I need to manually reset my DNS settings as my provider does not support DynDNS.

I have a similar setup to yours, Sky SR203 Router, Sky Q, 3 minis,  72 IoT devices running Tasmota and MQTT, and a few servers.

Pretty much all devices have fixed IP addresses, so its unlikely to be DNS.

I'm on FTTH so its very unlikely to be a WAN issue.

In my case disabling mDNS on the tasmota devices did improve matters, but its still not solved.

Sky have replaced the router and it still fails.

No devices are directly connected to the router apart from 2 GigE switches. 

The Sky 2.4G wifi is disabled as there is another problem where if you have a server (e.g. Home assistant) connected to the sky mesh via one of the Minis, it cannot respond to devices connected via another one so I have a completely separate wifi network.  i.e. its not the Sky wifi.

I doubt Sky has ever tested this router under the loads we are subjecting it to and also from experience solving load issues like this is not easy.

If anyone from Sky is reading this, perhaps ypu could confirm that you have tested the router under these sort of conditions, and if you'd like more from my set up, please DM me.

[GilesGodartBrown]

oops should have said

<snip>

Pretty much all devices have fixed IP addresses, so its unlikely to be DHCP.

</snip>

[TheBlackSheep]

I did think about upgrading the hub to the Max one with the extra bandwidth but that appears (judging by this forum that it's not got issues too (especially for SkyQ).  My system has actually worked fine since I sent my reply to you yesterday - I think it's listening...

I did think even dropping SkyQ (it's got very expensive) for NowTV or maybe even buying some static devices for Freeview (my wife doesn't watch much else so not sure I really need all the other channels).  I do watch a lot of the football hence NowTV maybe the only real solution.  I also have the landline connection (that's a bit hit and miss after a router reboot too - often requires a disconnect/reconnect through the router webinterface) - that's a bit more involved to replace.

I also think it's a volume thing - possibly a spike caused by a device but no idea how I find it. I used to make my own RJ45 cables up but I've tried to replace with high spec'd machine made cables to eliminate of my home-made dodgy cables. 

I need to start monitoring the network - possibly logging the router to a syslog server if that would give me any information - worth looking into (I'm not a network engineer)?

[TheBlackSheep]

..got issues too...!! -

why no edit on posts???