Bug Reports

Edit: Actually after logging some more entries, I am no longer sure that this happens at the time for the WAN DHCP renewal. Removed the snapshot from the syslog.
It hangs randomly, but it seems that it happens more and more often. Memory leak?

----

With m0n0wall 1.8.1 final release jan 15:
The m0n0wall seems to loose connection with all LAN clients when renewing the WAN DHCP.
All clients loose contact with the m0n0wall and you cannot reach the admin interface. m0n0wall has to be rebooted.

I connected external syslog to capture when it happened and after these entries, the system became non-responsive. Syslog entries are attached.
The actually assigned IP-adress is within the range.

The strange thing is that this started to happen after some time. The m0n0wall ran fine for many days and then this problem appeared.

Running on USB-stick on a PC unit.

I always start troubleshooting issues like this by making sure the platform is solid first.  I hate spending a bunch of time only to find out it was bad ram...   So start off by running memtest for a few passes to make sure you have a stable machine.  If that passes, give us the specifics on your hardware, like CPU, and nic, and how much memory.

Note that I have not mentioned the USB stick at all.  That is because once you boot, everything is in memory and the stick is irrelevant.

Hi Lee,

I started to suspect that it may be "something else", so I made more tests. Installed also a version of Pf Sense just to see what happened there.  Same problem after some time. So, as parts of these systems seem to be the same code, it could be caused by the same code, but it could also of course be hardware.

I switched back to m0n0wall now. Changed som settings for DHCP to activate IPV6 to see if it runs better. So far so good - however, I would be surprised if that would be the solution...
I will keep testing and if something clear comes up, I will of course post it.

Hi - Could this problem be caused by changes in the firewall rules?

Found a note on the reset state page about it being necessary to reset the firewall state after having done many changes. Is this a known "issue"?

Cheers.

The reason for that is that resetting states also resets connections.  Just making a rule does not, so an already going FTP download would not be cut off by a new rule blocking ftp.  It has nothing to do with this...

Hi,

This also did not make any difference.

Could instability in the WAN line cause this - meaning that it would make the LAN side to kick out its clients?

Not likely.  Try opening up an external web port and see if you can log in from outside while it is down.  It actually sounds to me like a hard crash.

Also, does the console respond?  Can you ping anything?