IPv6 DHCP-PD – radvd dies after interface reset - dhcpv6 does not reaquire addr
-
@ermal:
Can any of you run dhcp6c with -d -D options to have the logging of what is happening and paste that here?
Ugh, put pfSense back online and made that change. Not even pulling an address now.
dhcp6c[49192]: client6_init: skip opening control port
dhcp6c[49192]: client6_init: failed initialize control message authentication
dhcp6c[49192]: dhcp6_ctl_authinit: failed to open /usr/local/etc/dhcp6cctlkey: No such file or directorydhcp6c[83]: client6_send: transmit failed: Can't assign requested address
Same behavior as another user was reporting: http://www.broadbandreports.com/forum/r28333764-
–
I only had a short window to work with this, will try again tonight or tomorrow.
-
Works fine here:
root 30031 0.0 0.1 5780 1472 ?? INs 3:12PM 0:00.01 /usr/local/sbin/dhcp6c -d -D -c /var/etc/dhcp6c_wan.conf -p /var/run/dhcp6c_re0.pid re0As a side note, while preparing to make these edits, I noticed that I actually had two instances of dhcp6c running, with identical parameters and all. This is on a machine that was last rebooted earlier today, and before I had made any changes. I'm assuming that shouldn't normally happen?
-
Where does the additional logging go? /var/log/system.log?
-
Ok, I'm up and running with the -D option now.
I'm also seeing two instances:
root 43632 0.0 0.0 5780 1448 ?? Is 7:21AM 0:00.00 /usr/local/sbin/dhcp6c -d -D -c /var/etc/dhcp6c_wan.conf -p
root 43762 0.0 0.0 5780 1436 ?? INs 7:21AM 0:00.00 /usr/local/sbin/dhcp6c -d -D -c /var/etc/dhcp6c_wan.conf -p -
It only ran for about 50 minutes and then all IPv6 address was lost and radvd silently died. The only thing in the logs is:
Jun 6 08:32:16 dhcp6c[43632]: client6_timo: all information to be updated was canceled
Jun 6 08:31:29 dhcp6c[43632]: client6_timo: all information to be updated was canceled
Jun 6 08:13:00 dhcp6c[43632]: client6_timo: all information to be updated was canceled
Jun 6 08:11:11 dhcp6c[43632]: client6_timo: all information to be updated was canceledJun 6 07:22:03 radvd[26213]: version 1.9.1 started
Jun 6 07:21:56 radvd[57345]: Exiting, failed to read config file.
Jun 6 07:21:56 radvd[57345]: error parsing or activating the config file: /var/etc/radvd.conf
Jun 6 07:21:56 radvd[57345]: syntax error in /var/etc/radvd.conf, line 2:
Jun 6 07:21:56 radvd[57345]: version 1.9.1 started -
-
Part of the problem is that .conf is blank - i posted this before
Sure is.
I used to be able to pull IPv6 addressing and it would only go away after the lease expired after 4 days. Something has made the problem much worse since the mid-May build I was running.
I've been trying repeatedly. At times addressing will appear for a short time then go away. Most of the time I can't get any IPv6 addressing.
radvd.conf is blank.Jun 6 13:13:47 radvd[60869]: Exiting, failed to read config file.
Jun 6 13:13:47 radvd[60869]: error parsing or activating the config file: /var/etc/radvd.conf
Jun 6 13:13:47 radvd[60869]: syntax error in /var/etc/radvd.conf, line 2:
Jun 6 13:13:47 radvd[60869]: version 1.9.1 startedI have added the -D option ( /usr/local/sbin/dhcp6c -d -D -c ) and there is no additional output that I can find.
It's not a Comcast problem. If I fire up m0n0wall or Mikrotik ROS, it works fine, all the time.
2.1-RC0 (amd64)
built on Thu Jun 6 09:07:25 EDT 2013
FreeBSD 8.3-RELEASE-p8 -
Yup same here - I fired up m0n0wall, bing bang zoom everything working for ipv6. Now with pfsense if you do get an address its only there for a very short time.
I would think this should be a high priority issue - is there just something odd with the way comcast is going it that pfsense does not like? Or is this an issue across the board with other native ipv6 providers?
But I have to say that ipv6 is currently broken in pfsense - at least with comcast as isp in chicagoland area.
edit:
Be more than happy to let someone in to my pfsense install to do any sort of troubleshooting they would like to do.. Just PM me and can setup access. -
Yup same here - I fired up m0n0wall, bing bang zoom everything working for ipv6. Now with pfsense if you do get an address its only there for a very short time.
I would think this should be a high priority issue - is there just something odd with the way comcast is going it that pfsense does not like? Or is this an issue across the board with other native ipv6 providers?
But I have to say that ipv6 is currently broken in pfsense.
DHCPv6+PD may be broken, but that is not true of "ipv6" in general. My ALIX has a completely different issue in that it refuses to get DHCPv6+PD at bootup, but if I edit/save/apply the WAN, it comes up and stays up indefinitely. It gets DHCPv6+PD from my "main" home router, though, not an ISP.
My main home router does IPv6 via HE.net tunnels and it works fine.
-
DHCPv6+PD may be broken, but that is not true of "ipv6" in general. My ALIX has a completely different issue in that it refuses to get DHCPv6+PD at bootup, but if I edit/save/apply the WAN, it comes up and stays up indefinitely. It gets DHCPv6+PD from my "main" home router, though, not an ISP.
My main home router does IPv6 via HE.net tunnels and it works fine.
same here - but my ISP is RCS & RDS Romania - one of the biggest IPv6 ISP and no.1 in Ro - nationwide availability
-
Unfortunately, for me it's more or less the same.
But even the first offer is not accepted.Normally the link local address is used to aquire a lease.
I have a virtual IPv6 address that is used by haproxy to hide multiple servers behind a single IPv6 address.
now the DHCP6 client uses this ip as the interface IP so every offer is rejected with a XID mismatch.When I remove the virtual IP the initial lease is accepted but when the lease is expired the renew is denied.
/Andre
-
Interesting… I'm on Comcast, and my 2.1-RC0 box (June 4 snapshot) passed the four-day uptime mark today without losing IPv6 connectivity (and with rc.newwanip6 entries in the system log earlier today). Could this issue have been resolved?
-
I am still having issues.. Maybe I will try a clean install?
Just updated too
2.1-RC0 (i386)
built on Sun Jun 9 18:28:57 EDT 2013
FreeBSD 8.3-RELEASE-p8Jun 10 04:24:31 radvd[9391]: Exiting, failed to read config file.
Jun 10 04:24:31 radvd[9391]: error parsing or activating the config file: /var/etc/radvd.conf
Jun 10 04:24:31 radvd[9391]: syntax error in /var/etc/radvd.conf, line 2:
Jun 10 04:24:31 radvd[9391]: version 1.9.1 started
Jun 10 04:24:25 radvd[97175]: Exiting, failed to read config file.
Jun 10 04:24:25 radvd[97175]: error parsing or activating the config file: /var/etc/radvd.conf
Jun 10 04:24:25 radvd[97175]: syntax error in /var/etc/radvd.conf, line 2:
Jun 10 04:24:25 radvd[97175]: version 1.9.1 started
Jun 10 04:24:08 radvd[8125]: Exiting, failed to read config file.
Jun 10 04:24:08 radvd[8125]: error parsing or activating the config file: /var/etc/radvd.conf
Jun 10 04:24:08 radvd[8125]: syntax error in /var/etc/radvd.conf, line 2:
Jun 10 04:24:08 radvd[8125]: version 1.9.1 startedSo I get a ipv6 global and link local on wan.. But not seeing anything on lan side.. I have tried resetting back to none none for wan and lan, then back to dhcp6 and track.. I have rebooted, etc. Seems like some progress was having issues grabbing global ipv6 on the wan before, etc.
-
I'm on Concast, and was unable to get it to work on beta versions. I upgraded to the RC0 dated June 8 Saturday and it's been up and running since. Hasn't been 4 days yet so I don't know if it will stick. Not sure if it's fixed but it's working for the 1st time. I had been running M0n0wall because of the problems.
Interesting… I'm on Comcast, and my 2.1-RC0 box (June 4 snapshot) passed the four-day uptime mark today without losing IPv6 connectivity (and with rc.newwanip6 entries in the system log earlier today). Could this issue have been resolved?
-
I have put some more fixes for radvd wrong config error.
Please try a today snapshot. -
Nope. I did a fresh install of …
2.1-RC0 (amd64)
built on Tue Jun 11 06:04:26 EDT 2013
FreeBSD 8.3-RELEASE-p8... The 1st time it booted, it successfully got IPv6 addressing. I let it run a couple of hours and rebooted. Now there is nothing I can do to get it to pull IPv6 addressing again.
dhcp6c[68545]: dhcp6_ctl_authinit: failed to open /usr/local/etc/dhcp6cctlkey: No such file or directory
dhcp6c[68545]: client6_init: failed initialize control message authentication
dhcp6c[68545]: client6_init: skip opening control portAlso, radvd doesn't even show up as a stopped service, it does not appear on the services screen.
-
I tried the update from today and it broke the working IPv6.
-
I'm on comcast service. Zoom Modem is IPv6 enabled. PFsense, well can't keep a ipv6 address.
I'll stay at the latest beta that I have and update when there's a consensus that the IPv6 is working again.
-
Back on May 9th, I did a clean install of this version
2.1-BETA1 (i386)
built on Thu May 9 07:05:02 EDT 2013
FreeBSD 8.3-RELEASE-p8I was able to have IPv6 work for around 4 days before the lease from Comcast's DHCP servers renewed.
After that point, I was not able to get IPv6 working again.
What's interesting is if I do a full clean install, I still cannot get it to work. Its almost like something is left behind in some configuration somewhere that a full install does not remove.
I am looking to download and install the latest beta build, but wondering if the problem has been fixed yet. So, if someone has installed a newer build with some success, please pass that on.
–Brian
-
See my 6/10 post; I've had IPv6 connectivity on Comcast for more than four days straight with the June 4 snapshot. Haven't stayed on any one revision for long enough since.
