PFSense 2.0 Beta5 1/19 build system locks up

rpsmith

a few days ago I switched from my net5501/nanobsd to a Dell PC running I386 SMP full version. everything was running fine until I upgraded to "Sat Jan 29 01:09:59" and tried to connect to a remote site via IPsec site-site VPN. it immediately hard locked and had to be power cycled.

no packages installed
not using openvpn
no ftp traffic
single WAN and LAN
debug.pfftpproxy set to 1
running 4 Intel Pro 100 NICs

Edit: went back to "Fri Jan 28 14:23:00" and everything is back to normal. :)

Roy…

BlueMatt

Upgraded and have the same problem.
However, it seems to be triggered by several different triggers. Not sure what mine is yet, but it appears to have to do with the webConfigurator.

eri--

Yep know issue.
With the snapshot that will come late today everything has been fixed from what our testing shows.

mxx

Hi,

Just performed an upgrade to this snapshot (have been running one from ~26th).
Now pfsense doesn't boot anymore.

It didn't continue after displaying "Configuring Gateway Monitors".

I was continuously pinging the pfsense box from its lan side during boot and got exactly one reply back in between.
I then did a hard reset which led to that pfsense stopped loading further just one output line before configuring gateway monitors (if I remember correctly).
Silly me didn't take a screen capture of the first (regular) boot. I only have one after the reset :(

Screenshot-5.png_thumb

mxx

I don't know, maybe it's related to http://forum.pfsense.org/index.php/topic,32767.0.html since I have a few site-to-site ipsec tunnels configured and those remote hosts would try re-connecting all the time.. but I didn't want to hijack that thread..

BlueMatt

Similar problem, but it appears to boot properly if I disconnect my WAN interface prior to boot.
Also, even if boot is successful, I am not able to login to SSH (which previously worked).

MF

@ermal:

Yep know issue.
With the snapshot that will come late today everything has been fixed from what our testing shows.

if somebody is interessted in the symptoms

clean 1.2.3 install
updated via system_firmware_auto.php to Sat. 20110129 -> amd64

system locks up one CPU-core … seems to to nothing more

TUI tells

"Firmware upgrade in progress…"

reloading webconfigurrator gives```

Parse error: syntax error, unexpected '&', expecting T_VARIABLE or '$' in /etc/inc/interfaces.inc on line 1023


after "hard-reset" TUI tells

> "IPsec: Initialized Security Association Processing."

and locks up

mxx

You also happen to have any remote ipsec hosts connecting automatically?

Makkok

Same problem here, I'm wondering, last updates worse than ever… ???

eri--

I merged the topics because it really is the same issue.

I would really like the config files from the people that have installations with carp+openvpn.
You can send them privately with all private info removed.

As i said already it was a known issue and you have to wait for the snapshot of late today to test which
has commit https://rcs.pfsense.org/projects/pfsense-tools/repos/mainline/commits/ff8cebdea1046caacd263cab2b39053b629d3a83 in it.

mxx

@Makkok:

Same problem here, I'm wondering, last updates worse than ever… ???

Just bad luck ;)
Something like that didn't happen to me for a very long time.
I guess there's always a little chance of breaking something critical when developing such a huge project even more with snapshots being built automatically.
Of course this is very frustrating in a production environment. It's a good idea to have a backup machine to start up in such a case.

mxx

@ermal:

As i said already it was a known issue and you have to wait for the snapshot of late today to test which
has commit https://rcs.pfsense.org/projects/pfsense-tools/repos/mainline/commits/ff8cebdea1046caacd263cab2b39053b629d3a83 in it.

Hi ermal, is it ok to just manually apply that commit or would this break something as it may depend on something else that's changed?

clahti

so I have been on the Fri Jan 28 05:30:15 EST 2011 snap with the debug.pfftpproxy set to 1 and so far I have an uptime of 27+ hours (woo hoo). I am in the office today to build a 1.2.3 emergency system to slap in if the current 2x cluster goes down again, but as far as I can tell it seems stable at the moment. In reading the recent posts to this thread I gather that I should wait for Sunday snap since today's brings back problems for some people. ermal, you requested configs from people with this lockup issue, do you still need this? I can send to you if so, I am using CARP on two WAN and one LAN interface, but no ipsec or openvpn as our VPN traffic is port-forwarded to a different server inside the firewall.

eri--

if the hangs do not occur no i do not need the config anymore.

rpsmith

i386, Full, SMP, Sat Jan 29 15:43:14:

So far everything looks solid! No more Hard Locks when using IPsec tunnels!

Roy…

ivanmar

It seems this problem still exists.
Using built on Sat Jan 29 23:42:13 EST 2011 and PPTP VPN server and hat 2 locks during this day.
Now I will disable vpn server until this issue is resolved. Can not do onsite debugging because this is remote production firewall. I need to make it sable for now: Will downgrading to pre 19th help?

eri--

Just put a debug.pfftpproxy=1 and that will remove the hang.

ivanmar

I had 18h uptime, than again lock.
debug.pfftpproxy is set to 1
and enabled PPTP server

clahti

Running Fri Jan 28 05:30:15 EST 2011, I now have an uptime of 2 days, 23 hours. I am going to hold off upgrading to newer snapshots for the moment :).

clahti

Also as an aside, should we be putting debug.pfftpproxy=1 somewhere in a config file to survive a reboot?