I get a decent amount of interface errors, in and out, doesn't seem to affect anything noticeable at least. I'd like to know if I should work towards identifying the source and correcting it or if this is just expected due to some function of OPNsense I am unaware of. I'm not 100% sure how to identify what type of errors these are.
Quote from: Seimus on May 01, 2023, 05:55:28 PM
Are you using VLANs or any other virtual interfaces?
Arr those errors related to VLANs or other virtual interfaces?
Are you using a bridge which includes an unassigned interface?
Regards,
S.
I do have VLAN's but it doesn't seem directly correlated to just VLANs. See this screenshot (couldn't figure out how to post a table that looked decent).
Hm interesting,
I know that OPN has a bit of a problem with bridges or virtual interfaces as such (LAGG, VLANs, Tunnels), and it tends to create OUT or IN errors occasionally, mainly during bootup or startup of the interface itself, but that's a cosmetic bug and won't cause any issues.
Saying this, you should not see errors from the above-mentioned on any physical link or port.
If we omit the VLAN out errors you show based on what was mentioned above, what should maybe bother you are the physical LINKs (example: WAN IN errors). This basically means that a packet came "damaged," thus it was dropped.
Are those errors increasing?
If yes, on which interfaces?
Now the question is: what is connected to these physical links?
What do the interface counters on these devices say?
Can you check on OPN > Interfaces > Diagnostic > NETSTAT if you see any drops, for example?
Is your system suffering from possible interruptions?
If yes, is it possible for you to correlate input to the moment when an error occurs?
The problem of "error" as such is a debatable topic because it can be caused by various reasons (CRC, overrun, etc.). Naturally, from the GUI, you cannot deeply dive into specific counters. I believe the only way would be via CLI > SSH and looking at the interfaces from a BSD perspective.
Also, if I may, I want to comment on this.
QuoteI get a decent amount of interface errors, in and out, but they don't seem to affect anything noticeable, at least.
This is what is to be expected for a real "error" if you don't have excessive errors. Basically per network vendor specifications
(let's call it a gentleman's agreement).
Any error rate below 0.5% of the total traffic throughput per time is not visible from the perspective of performance for the user. There are, of course, some specific exceptions, but I don't believe you would use such services or applications.
Regards,
S.
The physical links are connected to multiple things, one connects to my 10g port on my ubiquity switch, one connects to my Xfinity modem, the other WAN connects to a LTE modem (error free but rarely used), and one connects directly to a NIC on my Proxmox host. The LAN connection and the primary WAN have the most errors, but as you said in the grand scheme of things it's a fraction of a percent of the packets.
Netstat shows the errors but no dropped packets on any interface. Any idea what the CLI commands are to view interface errors? I tried to search for it but came up blank.
This should post all data...
# ifinfo
Aren't you using Zenarmor? Netmap can also generate errors.
Cheers,
Franco
Quote from: franco on May 02, 2023, 07:05:59 AM
This should post all data...
# ifinfo
Aren't you using Zenarmor? Netmap can also generate errors.
Cheers,
Franco
Thanks Franco, yes I am running Zenarmor with the native netmap driver. So that's probably it. It feels like it's most likely cosmetic, I'm just naturally curious.