Visitor
•
4 Messages
Gateway fully cycling internet constantly (missing required option 24 and missing required option 82)
I moved into a new house on 6/10. I brought my equipment with me and kept the same plan. Immediately, my gateway began entering a full reboot cycle, happening 3 times a day on average. It was certainly very annoying but not crippling. Over a couple of months I spoke with tech support several times. I picked up a new gateway (TG4482A) a week ago since tech support was unable to understand or witness the issue. Since picking up this gateway, my internet fully cycles nearly every 5-10 minutes under use. Here are the primary scenarios that happen:
Scenario 1: all devices loose internet for a few minutes but still show a wireless connection (games will kick me, meetings fully disconnect, apps will show internet connection error messages).
Scenario 2: after scenario 1, my gateway will enter a full reboot cycle. Wifi is completely down, the lights on the gateway indicate a full reboot. This typically takes a full 10 minutes or so before things regain connection.
I really wish I was exaggerating with this issue, but it seriously happens constantly. It also seems very related to how much use the internet is getting. I hardly see bad logs from night hours, but during work (I work from home) and evenings it is constantly happening. Tech support confirmed that my connected/up time was average 15-16 hours per 24 hour day, so just over 50% uptime.
Maybe coincidence, but I have not been able to pull event logs from the gateway until yesterday when I got another set of new equipment. Since then, the logs are absolutely littered with critical logs of "missing option 24" or "missing option 82". These are logged within the same second that my gateway cycles its connection.
I have tried several options found in other forums (like moving to stateless dhcp stuff from stateful). Nothing has improved it. I have an advanced tech coming out today but I wanted to post here since the other threads are being closed for this issue.
bonus: it has even happened twice while typing this post
flatlander3
Problem Solver
•
1.5K Messages
3 years ago
The 82 and 24 errors during a boot wouldn't surprise me. option 82 is a fail talking to the switch directly upstream, I'm not sure if it's an up or downstream message in this context. It's a security feature to prevent spoofing. You'd need the firmware code to tell you what it's saying, and option 24 is just a dns-search path. If there's no link, both dhcp4 and dhcp6 fail, and so will the retries until the link comes up (obviously no dns search path at that point either).
I'd be more concerned about why the reboot happens, but none of these gateways keep a log history from before the boot as to why it tipped over in the first place. Signal/power issues will do it. I don't know if a dhcp fail will do it eventually or not. You'd need the firmware code to find out and Xfinity ain't saying, but line/cable problem is pretty likely.
For the router announcement/advertisement mode, I had to mess with it a bit on my stuff (BSD based) to get ipv6 working right. Most STD Options go something like this:
Unmanaged:
The firewall will send out RA packets and clients are directed to assign themselves IP addresses within the interface subnet using SLAAC. DHCPv6 is disabled in this mode.
Managed:
The firewall will send out RA packets and addresses will only be assigned to clients using DHCPv6.
Assisted:
The firewall will send out RA packets and addresses can be assigned to clients by DHCPv6 or SLAAC.
Stateless:
The firewall will send out RA packets and addresses can be assigned to clients by SLAAC while providing additional information such as DNS and NTP from DHCPv6.
Stateless doesn't work in my area, and from what I can tell, it's not using SLAAC just DHCPv6. I'm not running an internal DHCP6 server myself, Managed ought to be correct but isn't. Assisted seems to be the one to go with here, and the LAN port is set to "Track Interface". I don't know if you have the firmware option or not. In theory, since Xfinity approves the firmware, they ought to have tested it and selected the correct options for IPV6 for their implementation if set to auto, but perhaps not, or perhaps it varies by region. Again, Xfinity ain't saying.
The other part is the lease renew doesn't always work. Perhaps they get bombed with requests, or there's some malicious traffic messing with their stuff. Dunno. You're going to renew the DHCPv6 lease in 24 hours anyway, with a retry/renew in 12, so if there's an option, "Do not allow PD/Address release". You want get the new lease or renew before you dump the old one. You're going to overwrite it anyway. Seems to avoid stalls for me, at least so far.
(edited)
3