This morning, I noticed Wifi AND direct Ethernet connections had no internet access from my RBRE960 router. I checked the Xfinity app, it said the modem was online and fine. There was nothing in the modem log to indicate problems/errors. Both RBSs (RBSE960s) showed magenta LEDs. The switches seemed fine from their lights. So all signs pointed to the RBR not working. There were no lights on it, it looked normal.
Since there was no way to investigate further, I powered off the router, waited a full minute, then powered it back on. It came up, and all was working again - hop counts are 1 on both RBSs without intervention! WiFi is fine, direct connect Ethernet is working again too. RBSs showed Blue LEDs when they reconnected (they use Ethernet backhaul).
So besides the fact that the router stopped working and there was no way to find out why, I noticed the router's log file has been cleared from when I power cycled it. That's not normal, right? There was no firmware update - it's on V7.2.6.31_5.0.24 still.
In summary: - Router stopped working - no WiFi or direct internet access; - There really isn't any way to figure out what went wrong... - Power cycle fixed it - but its log was cleared.
How many days was it online before this happened this morning? Only thing I can think of is enable debug logging on the RBR at start up and if it happens again, see if you can save off the logs.
I noticed the router's log file has been cleared from when I power cycled it. That's not normal, right?
Yes, this is what happens when the router is restarted. It certainly is not what customers want to happen, but it is what happens.
There is a "debug" page which contains additional information about the router status, including an option to save information to a file. http://orbilogin.net/debug.htm If this happens again, before restarting the router, you could connect to the debug page and save the current contents of the file to a computer. Restarting the router clears this file as well as the regular log file.
Unfortunately, there appears to be no published information about how to interpret the debug file. With the 90 day period of 'complimentary support' being (literally) 90 days, most customers find that Netgear support wants them to purchase a support contract ($$$) before they can open a conversation.
How many days was it online before this happened this morning? 7+
Only thing I can think of is enable debug logging on the RBR at start up and if it happens again, see if you can save off the logs. ---> I couldn't access either the router admin web page or any running lan/wan capture logs when it happened. So that probably won't be useful.
@CrimpOn I could not get access to the debug page - I tried that. The network was completely down!
Ok, Ya hoping that NG can figure this out. Should not happen. Only thing I can think of is you've seen odd behaviors since you installed the XS series switch. 🤔
It could also be related to higher speeds - I used to be 1GB now it's 2GB. Probably not a ton of customers running at that speed on Comcast (which they have clearly had challenges with). Also, if they can't/don't test a ton of long-term issues, that would fit in - it ran perfectly for 7+ days.
Let's be honest here, there's potentially bunch of issues - pointing to firmware code that needs some serious walk-throughs, code audits, updates where later open source versions are available, etc.
What we have seen:
1- IPv6 issues - just Comcast maybe? Related to using Comcast DNS servers? (that has gone away by using Google DNS servers)
2 - Power cycle of the RBS then RBR, or just reboots will frequently require second power cycles of satellites (850 and 960 series) due to hop count being 0/blank so nothing can connect to it.
3 - system comes up unstable with 2 RBSs with ethernet connected backhaul - solution is often to let an RBS sync wirelessly then move it into position for wired backhaul
4 - adding a switch (XS508) and moving RBSs over to that switch, using their 2.5G ports, caused WiFi speeds to drop to 10% of their normal speeds starting 24 hours later.
5 - IPv6 requires about 125MB extra memory on the RBR - happens to some customers, most do not see this. Related to issue #1?
6 - @FURRYe38 haven't you seen memory and CPU usage creep/increase that doesn't go back down?
One last thought - over time it seems some users really have the issues - weird ones. You can identify them by browsing through the forum. I think I am in that camp too. My opinion, I think these are topology based - who the ISP is, how the backhaul network is set up, maybe the plan (speed), maybe other issues. It could also be mid-split enabled technology on Comcast, which is new and still being tweaked on their end. That started around here, early last fall.
For me it seems to be related to higher 1GB/2GB+ speeds, Comcast, and multiple wired back-haul.
ISP could be a big part of it, and even within the same ISP the networks are very different in different parts of the country - maybe even in different neighborhoods. For Comcast, they have mid-split enabled, low latency, DOCSIS 3, 3.1, 4, and different parts implemented to support these different technologies. That makes it hard to test against a particular ISP, there is no standard configuration.
If it was me at Netgear, I would probably review my firmware carefully, and develop a test version that logs/captures more info to determine why things fail and to correct it. Then ask some of these affected users if they would run a diagnostic version of FW, that captures/logs info to narrow down the problems. Maybe they have some work to do, to upgrade their firmware first, and if that doesn't make things more stable then try the diagnostic firmware. If they don't do something like this, it's hard to imagine they can find problems that only hit a small percentage of users after a number of days.
The creep I saw was on your 850 series when I had it online. Wireless saw no creep however after two RBS were ethernet connected, saw CPU usage creep upwards. One RBS saw no creep which was interesting. Got to around 30% CPU before we had a house hold brown out thus everything was reset and couldn't continue to test it until something happened. I had to take the system off line and couldn't continue testing. I do feel there maybe something with Ethernet connected RBS and switches at play that cause this odd behavior. Cabling as well. Yours was testing with out IPv6 here as my ISP doesn't support it natively so I don't think that's a factor for that configuration.
Though since your 960 series tested well into 10 plus days with IPv6 enabled and back then you were using GS series non managed switches that rated only 1Gb, now you have a multi gig switch in place and IPv6 enabled, could be anything. Wondering if it maybe possible that ISP IPv6 somehow caused the RBR to stop working or FW didn't handle something right.
Something to check as well, does the RBR case feel warmer or extra warm or hot? Something I keep under any router I have online, USB type 120mm fan 24.7 for additional cooling. Thought I don't think heat would have play a roll in what you saw. Just a suggestion here.
@FURRYe38 funny you should ask that...I have been frequently monitoring warmth on both the RBR and the XS508 for the last week. The XS feels like room temperature, almost cool; the RBR is close to room temperature - absolutely no heat issues at all.
