We currently have 2 FTTC connections, load balanced through a router using pfSense using PPPoE to the white openeach modems. Both been in for years running 47/7 - they are with two different providers.
Started noticing problems when we couldn't download files from certain sites, and generally poor performance. When looking into it further, this can occur on another site and whilst this site/server has an issue other websites are fine. Problems also arrise over SSH connections (so its not just HTTP) that stall and either timeout or eventually resume.
This seems to happen per desktop so if someone is stuck in one location, it may work for another machine.
Since we have two lines we pulled one and tested it with no luck, then swapped them and tried that. Problem seems to be on both lines so we started to look at our equipment.
Firstly we installed the most basic pfSense install we could and tried though this with no luck.
Then we got the ISP supplied routers (BT Hub etc) and a TPLink router out, plugged our computer directly into this and this didn't sort the problem (again, on both lines).
Tried many browsers for the http tests, and ssh through a few machines (openssh, putty etc) with the same issues, and even under other operating systems (Linux etc).
To avoid our network we installed a pc from scratch (instead of using our deployed image, we installed from a win7 disc directly), and plugged that computer directly into the BT router - none of this setup then touched any part of our pre-existing network - no luck.
After visits from 2 engineers (level 3 for the second one im told), we have replaced the RJ11 cable, faceplate was upgraded to the newest one, even tried an old filter in the master socket.
Modified MTU settings with no luck either. Can get baby jumbo packets 1508 running on one line and that doesn't help.
During all this speedtests are good, our ping is good, and the JDSU tests come back all good as well.
It's hard to explain this over and over again to staff because its not intermittent per se, we can replicate it whenever we want, and all they want to do is blame our equipement. Apparently one of the techs on the phone said to the engineer that the engineer needs to be frank with me that its our side....
I even got my Fedora installed laptop, plugged it straight into a modem, and used PPPoE from the laptop to connection and we still have the same issues.
We've even run our test setup on another power main (normally on a UPS controlled one) just incase something funky is going on there.
They also tried to blame it on our outgoing VPN routing but this was setup after our problem started for the purpose of having something useable in the interim - and it does work - if we redirect our traffic from our router over VPN to a remote VPS and out to the internet everything seems to work (bar a bit more latency due to the extra hop).
I'm completely lost as to what to do now - any suggestions?
Solved! Go to Solution.
It may be worth setting your routers to use either Google DNS or OpenDns.
*I want to add the true solution to this*
It was an OpenReach fault, after 2 weeks of problems some test they ran came back with a fault a few km away. Was fixed overnight.
So after being blamed over and over for this, and for the tech to tell the engineer to be "frank with the user" it was their fault all along.