Random Non-Responsiveness

[smajor]

Hi all, somewhere around 15.7.19, or my enabling DHCP, OPNsense has started randomly freezing. All interfaces go dead and I'm forced to power off and back on.  I run the CPU headless, so I don't see what might be happening on the screen. I've attached a display for "next time."

The interval can be completely random.  Several days to just hours apart.  Mostly the latter.

I thought maybe the older hardware I was on was having a problem, so I migrated to a newer box.  The exact same issues happen with it.

The web GUI doesn't detect a programming, but I submitted one via that mechanism anyhow in hopes that may show something. Are there any other logs available via the shell I can look at or submit?

[Andreas]

I Had this too
Just restart of the firewall helps or deactivating all dhcp server
Mostly I couldn't stop the dhcp service
It started it self new
Look up logs of the system
Mostly in my case the dhcp server died


Gesendet von iPhone mit Tapatalk

[smajor]

Really? DHCP is a core service, I'd hope it is stable. I didn't see anything over in the issues at Github for this. Maybe a dev will comment soon.

[smajor]

So, this morning I walk into no internet, no response on the LAN side, figure it's just a reboot again.  OPNsense comes up, still no internet. I can get to its console, no errors.  Reboot again. Check settings. Everything seems to be in order. Intranet is okay, WAN is configured, just no talking.

I pull the old OPNsense box I was using off the shelf, plug it in and everything works. I'm at a loss. Something isn't stable here, I'm going to try and investigate some more.

Edit: I've tried re-configuring both the LAN and WAN on the box, still no WAN. My WAN adapter is a USB dongle, I tried a different one, still no WAN.  I tried reverting to a previous configuration, still no WAN.  Status shows "up" LEDs on both ends of the adapters lit, just no traffic.

Whatever happened to this box probably has corrupted a service beyond my ability to locate. I will probably rebuild it again, this time sticking with 15.7.18. As I mentioned, the two recent changes I did were update to 15.7.19 and enable DHCP.  I'd much rather have an operable DHCP, so I'll test the other way first.

[franco]

We're investigating this, though this is so far out of our immediate scope to trigger this behaviour. We are looking for specific feedback and steps to reproduce with 15.7.20, it will most likely take a short cycle to 15.7.21 to get to the bottom of this and fix it for good. More on this tomorrow.

[smajor]

Thanks for the reply. I know it may not be of much help, but my usage for OPNsense right now is not too complex:

- WAN interface, statically assigned from my ISP.
- LAN interface, 10.0.1.1/24.
- DHCP 10-100 range and statically assigned addresses in 200-225 range.
- * URL lists, a mix of URL Table (IPs) and URL IP types aliased to firewall rules.
- About a dozen NAT enteries.

Not too exotic. Since rebuilding on the 21st it's been stable.

* I'm not using any aliases/url rules currently and am on 15.7.18.  I'm trying to test one thing at a time so I can at least tell you "when I added this, I saw it again."

[franco]

We've preliminary put some cache removals back in order to get to the bottom of this. The system is really slow and sluggish for at least a few people I have talked to. It looks like we've uncovered a deeper issue, no we go step by step, trying to reproduce which exact change caused this.

[smajor]

For what it is worth, I went all week with 15.7.18. No problems at all. I just updated to 15.7.20 with no other changes and I'll see how that goes.