Messages

Hi,

I run OPNSense 24.7.6. Last evening, IPv4 network connectivity broke. Restarting all services from the console menu did not help. A reboot was required.

This is what the General Log reveals:

Every 30 minutes, dhcp6c logs dhcp6c_script: RENEW on igb0 executing. igb0 is the WAN interface. Then, suddenly, radvd reports about 17 minutes later: sendmsg: No buffer space available. This message repeats every 5 seconds.

Pinging the LAN interface via IPv4(!) does not work. Neither can I reach any host on the LAN from OPNSense.

Only rebooting OPNSense brings operations back to normal. I had this once before with the previous version.

There are several reports of this issue in the forum but with little to no response nor solution. I am willing to contribute logs, configuration etc. to help solving the issue.

Kind regards,
Boris

I found the answer myself.

The MGMT port is connected to the board´s BMC and thus cannot appear as a network interface under OPNSense.

Here is the link to guide other people who come across the same topic: https://support.sophos.com/support/s/article/KBA-000009496?language=en_US

I must have enabled tmpfs for logging years ago to reduce the wearout on the SSD. I have disabled it under System | Settings | Miscellaneous. It requires a reboot, though, to become effective.

It just dawned to me: /var/log is on tmpfs and does not survive a power off/power on cycle.

Actually, they are not there. That's why I am asking. What you showed is what I expected but the directory only contains the current log file:

Code Select Expand

# ls -la /var/log/system/
total 61
drwx------   2 root wheel   128 Aug 22 14:01 .
drwxr-xr-x  12 root wheel   960 Aug 22 13:10 ..
lrwxr-x---   1 root wheel    35 Aug 22 14:01 latest.log -> /var/log/system/system_20240822.log
-rw-------   1 root wheel 61245 Aug 22 13:35 system_20240822.log

Dear all,

it randomly occurs that my OPNSense is suddenly not reachable any more. I am currently (still) on OPNsense 24.7.1-amd64 but had these issues say once every quarter for long.

It just occured ten minutes ago. I had to power the device down and back on. I read the thread with the same subject from today but it did not help. There is only the latest log showing entries since the last boot at 13:03 in /var/log/system. What I need are the logs from the previous run. Where have they gone? System | Settings | Logging, Tab "Local" is set to keep 31 log files.

Any help would be appreciated.

Best regards,
Boris

Hi,

I bought a used SOPHOS SG 430 Firewall Appliance and replaced the SOPHOS operating system by OPNSense 24.1. This runs all fine, including the LCD.

But I cannot use the MGMT ethernet port. It does not appear in the devices list (dmesg) and I also cannot access it via its default IP address 10.0.1.1 (https://10.0.1.1:4444). I did not find a mentioning of the management port in the web other than on the SOPHOS help page.

Any hints?

Kind regards
Boris

Thank you Franco, that worked.

I suggest to put this piece of information in the docs.

Kind regards,
Boris

Hello,

this OPNsense 22.1.8_1-amd64
FreeBSD 13.0-STABLE
OpenSSL 1.1.1o 3 May 2022.

This might sound stupid but I am missing the Services | Router Advertisement menu entry:

https://docs.opnsense.org/manual/radvd.html?highlight=advertisments#

radvd is happily running.

What can I do?

Kind regards,
Boris

It may be related to this: https://github.com/opnsense/core/issues/4121

I had to do something similar when OPNsense moved to logging targets...that was quite a while back.

Yes, looks like an related or even the same issue.

Export your conf file and grep for 31.2. If it's not there, it's not part of the configuration.

Thanks, Bart, that´s a remarkable trick!

Indeed, the rsyslog host is not in the configuration.

I moved the legacy* files away and rebooted. It worked, no more complaints about the non-existant syslog remote host.

Hello,

I am currently using OPNsense 21.1.7_1-amd64.

A few weeks ago I migrated the LAN of my OPNsense box from 192.168.31.0/24 to 192.168.11.0/24, thereby dropping the remote syslog server 192.168.31.2. Though, syslog-ng still tries to connect to 192.168.31.2. Since I did not find any trace of this setting in the GUI, I browsed through the config files and found /usr/logal/etc/syslog-ng.conf.d/legacy-remote.conf dated Apr 24 2020 still containing the entry.

Can I safely delete both legacy.conf and legacy-remote.conf in that directory? Is this a leftover from a previous update?

Kind regards
Boris

Further research led me to a Zotac ZBox CI329 which supports up to 16 GB RAM. I consider using a virtualized setup with Proxmox as host and OPNsense and FreeNAS as hosts - FreeNAS itself supports NFS and syncthing in a jail ("plugin"). FreeNAS is what I also use on the other end.

Only major issue is the lack of 802.11ac wireless adapters (even under FreeBSD 12.0 in an upcoming version of OPNsense) for access point mode.

I did some research regarding supported cards which confirms that the situation as described in the post cited by the-mk is basically unchanged.

Although there seems to be support for 802.11ac in the FreeBSD 12.0 kernel, I did not find any 802.11ac M.2 cards that can be used as access points. Not sure if there are any working adapters (drivers) for 802.11ac access point setup at all that can be used with a mini-PC (APU or ZBox CI329), be them M.2 cards or USB adapters.

For an APU you'd be just fine with a site2site VPN and rsync over ssh in the tunnel.

To my best knowledge, rsync is not capable of a two-way sync where changes, additions and deletions of files occur randomly on both ends and should be propagated to the other end. Apart from that, both sites will be part of a VPN.

The actual use case is having my home directory available at both sites locally to circumvent the issues related to NFS via WAN.