[SOLVED] Performance degradation after some uptime with no reboots

[dreamboat5299]

Just wanted to confirm if someone else observes this behavior. After some uptime (by some I mean at least 2-3 days) performance for some reason drops quite notably.

Not a huge issue personally, I can always reboot the router, but the speed drops from typical about 700 down to maybe 80 and ping rises from 20 to >100.

I've noticed it on at least last two versions, currently on OPNsense 24.7.8. Can not confirm or deny that it was there before.

Edit:
Using Intel NIC
os-cpu-microcode-amd is installed

[Seimus]

What HW do you actually have?
How is your LAN connected?
Ping to what has increased latency?

Regards,
S.

[dreamboat5299]

- HW: Fujitsu Futro S920.
- One WAN connection, all LAN connections through unmanaged switch directly to the opnsense machine with two interfaces, one of which is vlan.
- Connection tested with speedtest dot net, speedtest itself selects my isp as destination (i guess they host an instance).

Edit: I guess it's just something with my config, problem like this definitely would catch attention if it was somewhat widespread, but I don't see anyone else posting about this behavior. I'll just reinstall/reset my box and see if it will occur again.

[Seimus]

You say VLAN + Unmanaged switch.

So you are having VLAN on OPNsense, meaning you TAG?
Because if you are using VLANs and an unmanaged switch it may not be the best solution.

Also do you have a Shaper?
Shaper are CPU heavy, not sure what CPU is in that device (probably AMD G Series GX-415GA), but routing 700M on that CPU and having shaper can be taxing


You can try, when this happens again, disconnect the Switch and connect a PC directly to OPNsense to skip the unman switch to see if that makes difference.
Or if you run a shaper try to disable it if that makes any difference.

Also it would be good to observe the load and utilization of the CPU when this issue is happening.

Regards,
S.

[dreamboat5299]

Just to close this topic. I think I just raised a false alarm. In the end I have not reinstalled anything, running for several days now and issue does not seem to be there anymore. Could be that it was an issue on my ISP part. In any way, thanks for replies and explanations.

[tdan]

So, After updates I also experienced this, and once I rebooted the instance, speeds returned.  Recorded it happen on My speed monitor.  Prior to these updates, I haven't rebooted the instance for months.... It's on a bee link mini pc with gig fiber connection.