I also have a bunch of users running IPSEC connections into my network. I have have one of these guys: http://www.watchguard.com/products/xtm-8/overview.asp, and it actually does all the connection and routing logic. I understand that there are "workarounds" as you are proposing, but they are neither best practice or should even be considered given that the device given to us by Rogers should just work. Ultimately, like many people here, I really just need a modem. I have very specialized hardware for very specific functions. Yes, this is a residential connection, but that doesn't mean Rogers should be able to skimp on the fundementals right?
Overnight, the modem took the liberty of updating to the latest firmware - This morning at roughly 10:00, I stuck it back on bridge mode to see if the new firmware made a difference. I lost connectivity today at around 3:00.
I wonder if every single CGN3 exhibits this problem, or is it just a handful.
Yeah, my suggestion wont work for you. It woudl just cause more greif that solving. I noticed my internet glitched this morning when I was doing an RDP connection to my home PC. Tons of packet loss coming from home connection and speedtests were failing. An hour later it resolved itself. I haven't done a reset on my modem yet. Still trying to figure out of the issues I'm noticing is pure coincidence or not at this poitn since my modem is so new.
I did reach out to the social media team (twitter) and they did say that there isn't any known ongoing issues with the CGN3 and they are taking cases like these as one-offs. I was offered to have my modem replaced based on the information I provided them, but I said it's too early on to rule if the problem is with my modem. So I'm going to let the issue season first so I can truly understand the source of the problem.
I'll update this thread if it happens again and if I make any findings.
I spoked to tech support this afternoon - They instructed me to return the modem and get another one (CGN3), just to see what happens, so I just brought it back and popped it in. I noticed the firmware was back on the older version, then switched it to bridge mode. That was about 6:30'ish, so I'm guessing my Internet should cut out around midnight according to past experience. I'm enterntaining the possibility that this might be on the clock - Every 5.5 or 6 hours. I'll also report back if I find anything.
As I mentioned previously, I swapped my original CGN3 for another one yesterday, stuck it in bridge mode and let it go, fully expecting that my connection would drop out at some point during the night, as the original unit would only stay up for 5-6 hours at a time. Because of the numerous reports online regarding this issue I assumed that the issue was more deep and widespread.
To my suprise, I still had Internet this morning. It's been over 14 hours, and my connection has yet to hiccup. Could it be that the issue is confined to a certain batch of units? I'm still not completely convinced as of yet. Activity overnight for me is not as intense as it is during the day, so I'll spend the day testing and report back.
As for now, it seems as though switching out the modem for another one did it. So, if anyone is having this problem still, you migth want to consider switching the modems until you find a good one. Again, too soon to say, but I'll have more for you guys this evening after I've been testing for a day.
My connection just died 30 minutes ago. I had to do a complete powerdown and powerup to resolve.
The problem is that the internet connection isn't completely cut off, at least for me. If you have TCP connections already connected, they are almost always remain online. For example, my tweetbot twitter client for mac had no issues getting live updates being pushed to it.
Doing a traceroute going outbound will show spotty packet loss after the 2nd hop. You might have to hit refresh several times to get a website to load. DNS is extremely spotty due to the UDP nature of it.
Thinking it might be a buffer overflow of the ethernet port, I tried unplugging the ethernet connection and plugging it back in to see if that resolved the problem. Unfortunately it didn't. Now I'm starting to wonder if the modem is having problems tracking sessions because of a strange buffer overflow problem or something.
I can confirm that rebooting the modem instantly resolves the problem without issue. My next step if it happens again, and if I have the timing is to reset the CGN3 back into routing mode so I can put my home router into bridge mode to test the capability that way.
It's starting to look like Rogers will need to do some internal QA testing of this issue, and see if Hitron has any reports of these issues.
Just out of curiosity, what is your firmware version? My Internet, at least in bridge mode seems to last longer with the older firmware.
I too am having problems with the CGN3 modem, at least in Bridge mode. Last Thursday I traded in my old SMC Gateway (which was always in Bridge mode) for the CGN3. I brought it home connected it, and then changed to Bridge mode. Both Saturday and Sunday nights I started having issues - first just DNS problems in that I couldn't browse to web sites but some services, like an RDP like service, continued to work for a while. Eventually I tried a bunch of things but it was only fixed by rebooting the CGN3.
All was well on Monday, as far as I know - but I have now written a script that pings www.google.com and stores the results to a log file so I can keep an eye on things.
Tuesday at about 11am my service went down, came up a few minutes later, then it went down at about noon, came back a few minutes later, then it went down and stayed down around 1pm. I got home at 9:30pm, rebooted it and now it is working again.
I will try switching back to Gateway mode from Bridge mode.
FYI - to others monitoring your system - you can use a service like pingdom which will monitor a web page and let you know if it is down. Assuming that you have a web page of some sort running on your LAN this can tell you if your home internet goes down.