pfSense Plus Software Version 23.05 Release Candidate Now Available

liudger

Updated to 23.05 RC and most sites were unreachable. What to do? Can't find the exact cause. When I rebooted to 23.01 everything was fine again.

It looked like all ipv6 connections where terminated. But I don't have any weird rule that would cause this. Any suggestions where to look?

I have a PPPoE ipv4 connections and using Use IPv4 connectivity as parent interface for setting up an IPv6 connection.

markster

@liudger Have you determined yet if this is a DNS issue?
Can you ping any outside sites?

DefenderLLC

@markster If so, he needs this shirt:

markster

@defenderllc I want one of these.

DefenderLLC

@markster There's a YouTuber named Jeff Geerling that sells them. Here's the URL: https://redshirtjeff.com/.

liudger

@markster Sorry for the later reply. I can't ping ip certain ip's, but some websites work and some doesn't (doesn't make sense for me still). When I turn off pfblocker it doesn't make a difference.

This site is still reachable for example.

if I ping 8.8.8.8 I get this:

Request timeout for icmp_seq 1
92 bytes from pfsense.liudimation.com (10.0.2.1): Destination Host Unreachable
Vr HL TOS  Len   ID Flg  off TTL Pro  cks      Src      Dst
 4  5  00 5400 1a95   0 0000  3f  01 44ca 10.0.2.59  8.8.8.8

What is the next step to trouble shoot?

liudger

I found the issue. The gateway settings were changed when updating. And these automatic setting didn't work.

Changed it back to the ipv4 interface. So the issue that only ipv6 connections were working.
Also the gateway name is also suddenly all caps. This wasn't before. Maybe that is why it changed to other interface? Is this a bug?

stephenw10

Hmm, weird it looks like you only have one enabled v4 gateway there. I would expect 'automatic' to always use that. It may have lost it's default route somehow, setting a new default gateway would restore that.

DefenderLLC

Installed the next nightly RC build 23.05.r.20230513.0132. Is there a change log posted anywhere? Thanks.

liudger

@stephenw10 Yes that is what I would thought it would do. But also that it changed the gateway to automatic and that the name is suddenly all caps

free99

@stephenw10 Do you know when it will be uploaded to the repo?

stephenw10

Bug fixes. Platform recognition in Azure. Updating the repo data without packages present. PCAP page now handles interfaces that don't support VLANs correctly. Perhaps most exciting e1000 should now correctly handle VLAN0. Yay!

stephenw10

@liudger said in pfSense Plus Software Version 23.05 Release Candidate Now Available:

the name is suddenly all caps

Oh that's this: https://redmine.pfsense.org/issues/14057
Those should always have been all upper case and a bug allowed them not to be which would break upgrades from earlier versions.

stephenw10

@free99 said in pfSense Plus Software Version 23.05 Release Candidate Now Available:

@stephenw10 Do you know when it will be uploaded to the repo?

I assume the CE > Plus repo? That will be after 23.05 is released. It should be very soon after because we are now able to test that in advance so we know it works as expected.

liudger

@stephenw10
Than why is the other gateway (IPv6) not all caps?

DefenderLLC

Is there a change log anywhere for the pfSense+ RC builds? I believe this is the 4th RC release. Have not noticed much between them.

stephenw10

Good question. Perhaps the fix missed v6 gateways for some reason. Let me see if I can replicate...

stephenw10

Hmm, if you made a change to the v6 gateway it would have stored it in the config with the name at the time. Do you have gateway entries for either of those in your config file?

liudger

@stephenw10
This part of the config?

<gateways>
		<defaultgw4>KPN_SECOND_PPPOE</defaultgw4>
		<defaultgw6>KPN_Second_DHCP6</defaultgw6>
		<gateway_item>
			<interface>wan</interface>
			<gateway>dynamic</gateway>
			<name>KPN_Second_DHCP6</name>
			<weight>1</weight>
			<ipprotocol>inet6</ipprotocol>
			<descr><![CDATA[Interface KPN_Second_DHCP6 Gateway]]></descr>
			<monitor_disable></monitor_disable>
			<gw_down_kill_states></gw_down_kill_states>
			<monitor>2606:4700:4700::1112</monitor>
		</gateway_item>
	</gateways>

stephenw10

Yes, exactly! You only have an entry for the v6 gateway which is why it shows like that. The v4 gateway is still completely dynamically generated.

For most users the opposite it true. If they had ever made a change to gateway it will create an config entry for it and those should have been all upper case names. When this bug was introduced and gateways stopped being forced to all upper case causing issues for anything referencing it as upper case. This is now fixed. Unfortunately it appears you created the gateway entry while the bug was still in play. I would remove the gateway entry from the config, reboot, then recreate it from dynamically created gateway entry which will now be all upper case.

Steve