Balance One WAN connection drop

There have been a few threads about WAN connection dropping with certain modems. I had the problem with a Balance One and a Motorola cable modem (one of their newer cable modems). I would get several lost connections a day. I tried both PING and DNS health check, with no difference.

One thread suggested changing the Balance One WAN from Auto to 1Gb and un-check advertise speed. That did help on one unit, but on another the device would not commmunicate with the modem at all in that setting.

By experimenting I was able to completely fix the problem by using WAN speed 100Mb, and un-check advertise speed. Since changing from Auto to 100, I have not seen this problem for days. On Auto, I would see experience lost communication several times a day. My internet service isn’t 100Mb, so I’m not losing any speed by doing this, but such speeds are available. I would not be happy slowing down the router to communcation speed to accomodate this problem.

I’m running firmware 7.02.

Howdy Don. We tried each and everything you suggested but in a couple of cases we found that there was absolutely nothing that could be done to ensure a reliable connection. “100 and no advertising” did not solve the problem in at least two cases. We ended by by changing each and every modem to a non Motorola/Arris brand.
But, great news that that worked for you – thanks for sharing. I’ve been wondering about this issue since it has not been raised in a while.
It’s a sad day when a lower-end TP-Link/etc modem will work and a Motorola won’t! :face_with_raised_eyebrow:

1 Like

@Don_Ferrario, can I confirm putting a Gigabit switch in between will help?

1 Like

Rick and TK - apparently I spoke too soon. I ran two days with the WAN set to 100Mb with no problems, which is the first time I’d gone 24 hours without failure. Prior to that, the WAN dropped out several times a day. Unfortunately after two days running perfectly the problem resumed, and I lost connection half a dozen times in 24 hours, even with the speed set to 100Mb.

I now have a Gigabit switch bewteen the modem and the Balance One. The Balance WAN port is set to Auto. Working now, of course too soon to tell. I’ll keep you advised.

Hi Don. That’s totally consistent with what we’ve seen: Making such changes works for a looong time, days, minutes, or not at all. The “insert switch” suggestion made by @TK_Liew very often works – but sometimes does not. The only variable we’ve been able to identify is the darn Arris/Motorola modems.

Please keep us informed. :slightly_smiling_face:

Rick

The modems are the most popular in the world. They work fine with other routers. Peplink need to figure this out.

Going nuts with this. Putting a switch in between did not help, might make it worse. Its gone up and down half a dozen times in the last hour. I’m trying all combinations of port speed, advertise speed, MTU, DNS servers, etc. Gotta fix this thing before my wife makes me throw it away, and at the price of a Balance One I don’t want to do that!

I’m running firmware 7.02. Can you give me some earlier versions to try? Can I downgrade to something prior to 7?

Does the problem go away if you disable the health check? What is captured in the event log?

1 Like

Thought of that, but the problem does not go away without health check. The connection drops are real. When health check is disabled, I know the drops happen because our VPN connections drop, then later reconnect.

A few hours ago I rebooted everything, and have the WAN port set for 100Mb. Its been stable for four hours but of course its done that before.

One interesting thing is that if I set the WAN port to 1gb (not auto), the router will not communicate with the modem even if there is a switch in between. It just sits there spinning at acquiring IP. Change it back to 100Mb, and things work at least for a while.

@Don_Ferrario, since the problem persists after putting a switch in between. Then the problem is not related to the port compatibility issue. I suggest open ticket for us to investigate.

Thanks for your patience.

1 Like

I have made some progress on this issue. I installed a switch between the Balance One WAN port, and the modem. That made no difference at all. I still got between 3 and 10 dropped connections per day.

I started experimenting with the various WAN settings in the Balance. Port speed, MTU, advertise, etc. None of it made any difference. Sometimes it would work well for a day but then the problem returned.

The locations where I have this installed has only one internet connection. When I install in that situation I set the upload/downoad bandwidth to 1Gbs. As far as I know, those settings are only used for load balancing, and since there is only one WAN port in use, there is nothing to balance.

I tried lowering the bandwidth settings. Every time I lowered the setting, the link became more reliable. I ultimately settled on 35Mbs upload/download, and the link is stable. It has gone down only twice in week, and its entirely possible the internet provider was at fault for those. Two times in a week is a whole lot better than ten times a day. This is an Xfinity cable internet, and their stated max provided speed is 35 down, 5 up, so thats why I settled on 35Mbs.

In my opinion this is clearly a flow control problem. It should not be necessary to throttle the WAN port this way. The port should send/receive data at 1Gbs speed, and use normal flow control as needed. I don’t know if this is a firmware issue, or a hardware flaw in the Balance One.

I have another Balance One with a single cable internet connection (Time Warner/Spectrum). That one is set for 1Gbs bandwidth. I have no issues with that installation.

1 Like

I’m having the exact same issue with my Balance One and I have confirmed that it is not isolated to one router. My original B1 was having the random disconnect issue and one morning I came in to the office and the router was completely dead. Did a RMA and in the meantime bought another B1 from Amazon to run while the RMA processed. Same issue on this new router here and there it would drop WAN. Nothing in the event log that indicates issue.

Once I got my RMA replacement in the mail I hooked it up - did a fresh config to start off with and I’m having the same issue again. Prior to this I have had zero issues with the same exact network setup for over two years.

Yesterday I did tray and downgrade the firmware thinking maybe 7.0x was the issue and about 3-4hrs after leaving work I figured out the network was down again. Every time when I come back in the office I cannot even login to the web admin console I have to unplug the router, plug it back in, and voila everything is back to normal. Again - nothing in the event log. Same issue with three different routers on 6.3.4 and 7.0.1 and 7.0.2

1 Like

Jake, just to test my theory, I suggest you reduce the bandwidth settings on the WAN page, to something below what you are being provided by the internet provider. See if that helps.

I know from experience, at least with prior firmware, that the bandwidth settings will throttle the data transfer. With only one WAN, it should not be necessary but at least for now if it fixes the problem…

All I can tell you is that if Peplink would exchange the two Balance One that I have, for B210, I’d take it in a heartbeat. I do not have this issue with any other models.

1 Like

Yeah it’s kind of frustrating since it’s not an issue showing in syslog or anywhere else. Router just effectively takes a dump and stops working at random. Only thing that will fix it is unplug and plug back in. Can’t login on LAN when it’s “dead”

I’m trying your bandwidth settings to see if that helps.

1 Like

@jakebdb56

Do you open a support ticket to allow support team to check ?

For @Don_Ferrario case, issue identified due to health check server issue and this had been confirmed by Don and the problem resolved by changing using different health check servers.

1 Like

Further explanation in sitloongs post. I had a similar problem with a Balance One behind a cable modem. I tried both a Netgear and Arris modem. I tried inserting a switch between the Balance and the modem. The problem was reduced by not eliminated by lowering the bandwidth speed settings.

The health check had been using the ISP’s DNS server. Its Comcast Xfinity, they should have good equipment. By changing the health check to a public DNS server 8.8.8.8 or 9.9.9.9, the problem has been eliminated. I didn’t actually change the DNS, I only changed the server being used for health check. I can only assume that when the ISP’s DNS server gets busy, it stops responding to heal checks.

1 Like

@Don_Ferrario

Thank for the details explanation :+1:

1 Like

Are you guys talking about (Health Check Failed) or (Link Down) errors?

The former seems to be a problem with DNS servers, but the (Link Down) problem is the one that I only see with Motorola modems.

In my case, with the (Link Down) error, the link always comes back online in about 3 seconds. Annoying but not fatal. It only happens with an ARRIS 1672 modem, but never happens with a Netgear CM600.

I think these are two separate issues and this thread is confusing because I think people are not talking about the same thing…

1 Like

At least in my case the problem has always been health check. Whether I was using DNS or ping, I had still been using the ISP’s DNS servers. Changed to public DNS servers for ping or DNS request, and problem has gone away.

Only time I’ve seen link down is when something really was off line.

Mine doesn’t bring up any errors in the log which is frustrating because I don’t have an easy method to find out what’s causing the problem.

I did change the bandwidth settings for the WAN’s and it’s been fine for a few days now without any issue, but we will see how it goes through the next week. knocks on wood

With what I am seeing - something happens in the Balance One and literally everything stops in/out without any error message. Only thing that fixes it is a power cycle. It comes back online without any logs of an issue.

Modem is a Netgear CG3000DCR