Messages

I agree with everyone... I've been in technology for my whole career so I always start troubleshooting with "what changed". It was nothing on my end for sure. Which left: daylight savings change and cable modem that I could think of.

So, yesterday was a trial of patience with the WAN going down every hour (on the hour... 3600 renewal) A reboot of the VM OR the cable modem would fix it temporarily.  And then suddenly (again with the "suddenly") it would not get a WAN address at all. No amount of rebooting the vm or the cable modem changed things. It was late and since it already didn't work, I did a factory reset on the modem. This resulted in immediately working again and stayed that way for ~18 hours. Alas, it died an hour ago and none of the previous fixes would work.  DHCP request broadcast never got a response. I placed a call to the ISP and they "sent a reactivation code" which started things back up. It's working now but i'm dubious that it will continue. I mean, it was working fine for months and then something made it start this cycle. Crossing fingers but expecting to be buying a new cable modem soon.

Thanks for all of the suggestions.

It seems to be locking up on lease renewal which is every 60 minutes. when it does, there is no indication in the log (that I can tell) except that no VPN gateway is available (since the VPN runs over the WAN).

I would like to think its an ISP IP conflict but whenever it goes down I and I reboot the opnsense vm, I get a new WAN IP. It will work just fine for an hour. I've now had dozens of WAN IPs; they can't all be conflicted.

1) Lease Renews - WAN dies
2) Reboot OpnSense - get new WAN IP
3) It all works for an hour
4) Repeat

this just started on friday with no changes to OpnSense at that point and my ISP says no change on their end.

I've been running for several months on 18.7 with no issues and suddenly (as if I'm in a mystery novel) two days ago I started getting no internet access every few hours. WIFI and LAN are working fine. WAN shows down. A reboot of the opnsense vm fixes it for another couple of hours.

After few days of this I backed up my VM and then upgraded to 19.1.4 as a trial. Same issue.

I noticed this morning (after a reboot) that the WAN IP address had changed when it came back up. After another two hours... WAN is down again.

I tried powering down the cable modem overnight but that has not solved anything.

I'm technical but not versed in troubleshooting this. Any thoughts on where to start looking? cable modem\ISP? Opnsense issue? Both?

And why oh why during March Madness?!?! <argh>

Thank you to any who can point me in the right direction.

@bimmerdriver

Be thankful yours worked so easily!

I wonder if its hardware, host OS or something different that has caused issues for myself and others.

I have E5620s  ,Broadcom(Qlogic) NICS and am on HyperV 2016 Core.

I'm not certain that disabling integration services had anything to do with success. I may test more on that later, but after many frustrating hours of no WAN connection, I reverted to as simple as a setup as possible to get my Gen1 VM to work. Then I just repeated that on the Gen 2 VM.

Now that I have a setup that is working again, I would like to try to install with integration service and SRVIO.

Still wondering why I had to clone the WAN-side physical NIC MAC after the host rebuild (UEFI\GPT) when it didn't need it previously (BIOS\MBR). I reset my cable modem several times and tried forcing a lease renewal on the physical NIC during the whole ordeal so I'm pretty sure it wasn't a lease issue.

Dunno... hopefully someone cannot benefit from my struggles.

[Finding #1:]

Wanted to take a moment to share some of my trials with 18.7, UEFI and HyperV Gen2; especially how I got it to work...

My setup: PowerEdge R710 running Hyper V Core 2016. 1 NIC sits on the WAN and 1 sits on the LAN. My plan was to setup an external Vswitch for each NIC and do my routing and VPN through an OpnSense guest.

Initially, I could never install OpnSense on a Gen2 VM. It would hang in various places and no amount of font resizing or console mode changes, helped. :( So I installed on a Gen1 and it went without a hitch. This was my first time with OpnSense and my first time with HyperV so I was very happy with this initial success. :)

I realized the other day when I was attempting to add some storage that my bios was set to bios mode and not to UEFI. That will not do as every homelab needs > 2TB. :o So I exported the VM and reinstalled HyperV on the host in UEFI mode. (by the way, MBR2GPT doesn't work on poweredge servers... due to raid???) This part went without a hitch.

I then restored my OpnSense VM to the fresh host and fired it up expecting it to work immediately. Ummm. not so much. The Vswitches were built the same, everything else was configured the same, but OpnSense could not see the WAN. It saw the VNIC interfaces; hn0 and hn1 were present but no WAN connection. :'( Screwed with it for >12 hours until I finally hit upon something. On a desperate whim, I set my VNIC MAC to static and copied in the MAC of the physical NIC; it magically started working. ???

Finding #1: At least on my PowerEdge in UEFI mode, dynamic MACs and arbitrary static MACs won't work on the WAN side. (I hope someone can explain this) (Recall this worked in MBR\BIOS mode... just not UEFI)

So now I have my Gen1 VM working again. Great. Kids are happy that Netflix and PS4 are working. ::) Feeling like the hero, I decided to see if this would help with the similar issues I've read about on the Gen2.  So I created a Gen2 VM and booted from the installer CD using the same MAC clone workaround. Immediately got a WAN and LAN response. YEA!   But when I tried to install I hit the same freezeup at the "guided install". So, after a few trials, I booted again from the CD, but once it reached the login, I closed the console and opened up SSH. I performed the rest of the install from SSH without a hitch. Following the standard post-install reboot I had a Gen2 VM working. I pulled my config over from the Gen1 and now the Gen2 is up and running with all my setup. ;D

Finding #2: Don't try to complete the install from the console on a Gen2 VM in HyperV (Neither EFI nor VGA modes worked for me). Run the first phase (up to the initial login) on the console and then complete it over SSH.

My configuration:

• Gen2 VM
• No SCSI, All IDE.
• No integrations services. (Though I may try to add back "Shutdown" and "Time Sync")
• Static MACs cloned from the NICs to the VNICs.
• Network hardware accelerations all turned off
• Boot from install CD to login prompt and then close the console
• Complete the install over SSH

Sorry for the saga... too many reboots and not enough sleep. 8)

Hope this helps someone and I really hope someone can explain any of it.
 

 

I have setup my opnsense box on 18.7 with multiple OpenVPN connections. I have a gateway group setup as load-balancing with these two gateways only.

If I have both of the opnvpn clients up at the same time, the speed (using speed test for instance or just trying to access an internal web page) is incredibly painfully slow.  stopping either one of the openvpn clients restores full speed. restart it and immediate slowness.

Has anyone experienced this?

I am going to try testing with two different vpn services later to see if it could be service related, but I doubt it since I had it working like this in another routing setup recently.

In 19.1 I am unable to "Save" anything on the Routes page. Click the Save button and nothing happens.

Thoughts?