DNS Forwarder stops working in pfSense 2.0.1-Release
-
Not sure, was hoping it would be this week, still waiting on some other info to come in
-
Nice! :)
Thx mate!
-
Anything in the System Logs when that happens?
I haven't seen that before (or heard of it) that I can recall and I've worked with hundreds of installs in all kinds of situations.
2.0.2 will be out before too long here so it's probably best to hold off on reporting such a thing until trying it with the latest version (once it's out).
It seems like dnsmasq died at 03:54:27
It restarted at 08:33:16, that's the time I entered the webconfigurator and just saved the configuration as it was (no changes made).Does this give any insight to you?
I've got this:
May 24 02:09:25 etna dnsmasq[40901]: read /etc/hosts - 15 addresses
May 24 02:09:25 etna dnsmasq[40901]: read /etc/hosts - 15 addresses
May 24 03:05:04 etna kernel: pid 59534 (ps), uid 0: exited on signal 8
May 24 03:27:44 etna kernel: pid 33445 (awk), uid 0: exited on signal 8
May 24 03:54:27 etna kernel: pid 40901 (dnsmasq), uid 65534: exited on signal 8
May 24 04:06:31 etna kernel: pid 55056 (awk), uid 0: exited on signal 8
May 24 04:07:35 etna kernel: pid 11563 (awk), uid 0: exited on signal 8
May 24 04:08:38 etna dhcpleases: Could not deliver signal HUP to process because its pidfile does not exist, No such process.
May 24 04:08:38 etna dhcpleases: Could not deliver signal HUP to process because its pidfile does not exist, No such process.
May 24 04:50:40 etna kernel: pid 31474 (rrdtool), uid 0: exited on signal 8
May 24 06:07:52 etna dhcpleases: Could not deliver signal HUP to process because its pidfile does not exist, No such process.
May 24 06:07:52 etna dhcpleases: Could not deliver signal HUP to process because its pidfile does not exist, No such process.
May 24 06:29:33 etna kernel: pid 21544 (awk), uid 0: exited on signal 8
May 24 06:32:47 etna kernel: pid 20891 (awk), uid 0: exited on signal 8
May 24 06:41:23 etna kernel: pid 52679 (awk), uid 0: exited on signal 8
May 24 08:01:49 etna kernel: pid 32847 (awk), uid 0: exited on signal 8
May 24 08:07:05 etna dhcpleases: Could not deliver signal HUP to process because its pidfile does not exist, No such process.
May 24 08:07:05 etna dhcpleases: Could not deliver signal HUP to process because its pidfile does not exist, No such process.
May 24 08:31:48 etna php: /index.php: Successful webConfigurator login for user 'admin' from 192.168.69.18
May 24 08:31:48 etna php: /index.php: Successful webConfigurator login for user 'admin' from 192.168.69.18
May 24 08:33:14 etna check_reload_status: Syncing firewall
May 24 08:33:15 etna dhcpleases: Could not deliver signal HUP to process because its pidfile does not exist, No such process.
May 24 08:33:16 etna dnsmasq[46339]: started, version 2.55 cachesize 10000
May 24 08:33:16 etna dnsmasq[46339]: compile time options: IPv6 GNU-getopt no-DB us I18N DHCP TFTP
May 24 08:33:16 etna dnsmasq[46339]: reading /etc/resolv.conf
May 24 08:33:16 etna dnsmasq[46339]: using nameserver 208.67.220.220#53
May 24 08:33:16 etna dnsmasq[46339]: using nameserver 208.67.222.222#53
May 24 08:33:16 etna dnsmasq[46339]: ignoring nameserver 127.0.0.1 - local interface
May 24 08:33:16 etna dnsmasq[46339]: ignoring nameserver 127.0.0.1 - local interface
May 24 08:33:16 etna dnsmasq[46339]: read /etc/hosts - 15 addresses -
Signal 8 is a Floating Point Exception.
I would seriously doubt your hardware.
Especially since many things are dying, not just dnsmasq.
-
Signal 8 is a Floating Point Exception.
I would seriously doubt your hardware.
Especially since many things are dying, not just dnsmasq.
I've been testing the hardware extensively before installing pfSense. If this keeps happening I'll try changing memory module, but I am not prone to think this is a hardware issueā¦ Floating point exception... some strange CPU/FPU related thing?
Do you know if there is any way to monitor the thermal levels of the mainboard? This is a Nortel Contivity appliance that I retrofitted with pfSense, both fans are working ok, and temperatures look OK at BIOS settings, but I can't check them while running.
-
I've been testing the hardware extensively before installing pfSense. If this keeps happening I'll try changing memory module, but I am not prone to think this is a hardware issueā¦ Floating point exception... some strange CPU/FPU related thing?
Possibly a memory fault that turns an instruction into a bizarre floating point instruction.
-
Anything in the System Logs when that happens?
I haven't seen that before (or heard of it) that I can recall and I've worked with hundreds of installs in all kinds of situations.
2.0.2 will be out before too long here so it's probably best to hold off on reporting such a thing until trying it with the latest version (once it's out).
It seems like dnsmasq died at 03:54:27
May 25 03:46:04 kernel: pid 42320 (dnsmasq), uid 65534: exited on signal 8
May 25 03:50:08 dnsmasq[55898]: started, version 2.55 cachesize 10000Almost the same time, one day laterā¦ It's been working ok all day long... restarted it now. Let's see when it dies again...
-
Any clues in the time it dies every night??? its about the same timeā¦
-
Any clues in the time it dies every night??? its about the same timeā¦
That's at most a coincidence, check all the signal 8s spread all over numerous times on multiple completely unrelated processes. That's pretty much guaranteed to be a hardware problem.
-
The same problem here with Nortel Contivity and pfsense 1.2.3. It happens 1-3 times a month. I've seen people with monowall complaining about the same issue. It must be something with hardware platform and freebsd incompatibility. Any chance about running script/cron job to check and restart dnsmasq service when it happens? Every minute or maybe every 5 minutes? That would do it for me. Crontab is not good place to add that because it might be overwritten. How about adding it inside /var/cron? I'm OK with Linux, but I would like some input from people with pfsense/freebsd experience. And yes, I still didn't upgrade to pfsense 2.0.1. Actually, it's on one of my customers machines and it looks like upgrading to 2.0.1 will not resolve this issue. And 1.2.3 works just fine for them for now.
