Lightsquid service not starting at boot time

cwagz

I use pppoe with ipv6.
Guys with problems, do you also use pppoe?

I am on DHCP.

maverick_slo

I think it`s related to this:
https://github.com/pfsense/pfsense/commit/80d3cf1cdd79517cad7676813641d2e0fe73b445
https://redmine.pfsense.org/issues/5952

This is definitly a bug in my opinion.

maverick_slo

Hi!

I also have:

/rc.start_packages: Skipping STARTing packages process because previous/another instance is already running

And a lot of those (I don`t know if it is connected):

Mar 14 21:07:45 	php-cgi 		rc.bootup: The command '/sbin/route change -inet6 default fe80::2c1:64ff:fed8:3f00' returned exit code '1', the output was 'route: writing to routing socket: Network is unreachable route: writing to routing socket: Network is unreachable change net default: gateway fe80::2c1:64ff:fed8:3f00 fib 0: Network is unreachable'

/rc.filter_configure_sync: The command '/sbin/route change -inet6 default fe80::2c1:64ff:fed8:3f00' returned exit code '1', the output was 'route: writing to routing socket: Network is unreachable route: writing to routing socket: Network is unreachable change net default: gateway fe80::2c1:64ff:fed8:3f00 fib 0: Network is unreachable'

/rc.start_packages: The command '/sbin/route change -inet6 default fe80::2c1:64ff:fed8:3f00' returned exit code '1', the output was 'route: writing to routing socket: Network is unreachable route: writing to routing socket: Network is unreachable change net default: gateway fe80::2c1:64ff:fed8:3f00 fib 0: Network is unreachable'

/rc.dyndns.update: The command '/sbin/route change -inet6 default fe80::2c1:64ff:fed8:3f00' returned exit code '1', the output was 'route: writing to routing socket: Network is unreachable route: writing to routing socket: Network is unreachable change net default: gateway fe80::2c1:64ff:fed8:3f00 fib 0: Network is unreachable'

maverick_slo

What fixed it was editing /etc/rc.start_packages and changing:


FROM:
if (time() - $stat['mtime'] >= 90) {

TO:
if (time() - $stat['mtime'] >= 5) {

So it does only wait 5 seconds and then starts packages.
It`s a dirty hack and will work only until next update or gitsync.

I`m still very much interested in resolving this without hacks like this.

jimp

Sounds like this actually has nothing to do with lightsquid but is instead a general package startup problem. Feel free to open a bug entry on redmine.pfsense.org with what you've found so far. I would not emphasize lightsquid on the redmine report since it happens to other packages as well for some people (e.g. Avahi)

maverick_slo

Here we go: https://redmine.pfsense.org/issues/6001

maverick_slo

I`ve updated https://redmine.pfsense.org/issues/6001#change-25756 with new findings.

Snort takes really long time to start and this interrupts other packages.

Is there a way to alter the package startup order?

maverick_slo

OK final post here as this won`t change in near future :)

I found a workaround.

Edit config.xml and move lightsquid to top of packages.
The XML tag is: <package>So lightsquid will start first and snort last in my case.</package>

cwagz

This mostly worked for me as well. I moved light squid first and Avahi second. Light squid is starting now at reboot but Avahi still requires a manual restart.

cwagz

Seems like a similar error I was seeing in previous versions while running in hyper-v. Some type of timing issue. Should this be a new topic? I was thinking it is all related.

Mar 19 21:58:46 avahi-daemon 54193 dbus_bus_get_private(): Failed to connect to socket /var/run/dbus/system_bus_socket: No such file or directory
Mar 19 21:58:46

jimp

Anyone that was hitting this problem, update to the latest snapshot and see if the problem persists. A fix was put in yesterday that should address the problem.

maverick_slo

I just updated my ticket.
It`s OK and others report that Avahi is starting too.

Thanks!