Jake,
Am 18.06.2014 20:28, schrieb Sallee, Jake:
> Apparently Apple has introduced a new feature for iOS8 that generates
> a new MAC for the phone every time the user connects to wifi.
if I read the press reports correctly, the phone will only use the
random MAC for scanning. Once it actually jo
Hi.
According to the Cluster Guide, mysql is supposed to bind to the
management-ip and the various host settings in pf{,config}.conf should
be '127.0.0.1'. But this doesn't work as mysql doesn't listen on
127.0.0.1.
I'm assuming this setting should be the management IP of the cluster as
well,
Hi Louis,
Am 24.03.2016 19:51, schrieb Louis Munro:
> The trick is that you don’t connect directly to Mysql, but rather to
> HAproxy which will transparently connect PF to the database.
Thanks! I knew I was missing something basic...
Regards
felix
--
Hi.
We want to lock our wired node registrations to the switch and
switchport they were registered from to prevent users from cloning other
users' MAC (which has happened in the past). I had hoped to be able to
do this with vlan filters similar to what was suggested in
http://permalink.gmane.o
Hi everyone.
Am 25.04.2016 20:51, schrieb Till Wimmer:
> sadly, it was not the timezone setting. After runing for almost a day,
> the problem did come back. Average sytem load is 1.34, Dhcpd is running
> 100% of one CPU.
We're seeing the same behavior, but only on one of the two nodes in the
clu
Hey.
Am 25.04.2016 21:40, schrieb Derek Wuelfrath:
>> We're seeing the same behavior, but only on one of the two nodes in
>> the
>> cluster.
>
> Running RHEL (includes CentOS) or Debian ?
Debian wheezy.
An strace on the 100% cpu dhcpd looks like this:
select(30, [4 5 7 8 9 11 12 13 15 22 23 24
Hey.
Am 11.07.2016 23:45, schrieb Louis Munro:
> We have had a look at Kea, but it still seemed a little bit rough
> around the edges…
Afaik Kea still does not support failover (and depending on network
topology, using a HA database backend can cause IP conflicts in a
split-brain situation).
Hey.
Am 18.08.2016 22:01, schrieb Jan-Patrick Perisse:
> In PF ZEN V6.1.1.
> It was taking 42 seconds for the mysql query that shows nodes to
> complete.
> I added an index for radacct.callingstationid and it went down to less
> than 1 sec.
This has been fixed in Packetfence 6.2, see
https://pa
Greg,
Am 29.09.2016 17:35, schrieb Greg Harewood:
> WHY DOES THE pf.conf file have local node information in it, but is
> synced as a shared file?
>
> I'm thinking of the interface information primarily in mine,
This is overridden in cluster.conf, see
https://packetfence.org/doc/Pac
