Kresd - enabling the http module

I found someone else trying this:

But I don’t have the problem they faced. I think.

I just found this:

And would like to have that available.

So yes, I have a:

# cat /etc/kresd/custom.conf 

and I restart:

# /etc/init.d/resolver restart
Called /etc/init.d/kresd stop
set dhcp script
Called /etc/init.d/kresd start
set dhcp script
Called /etc/resolver/

So I expect it to be listening on port 8453 but:

netstat -tulpen | grep 8453

produces nothing. And a full list of kresd ports:

# netstat -tulpen | grep kresd
tcp        0      0    *               LISTEN      10735/kresd
tcp        0      0 :::53                   :::*                    LISTEN      10735/kresd
udp        0      0 *                           10735/kresd
udp        0      0 *                           10735/kresd
udp        0      0    *                           10735/kresd
udp        0      0 *                           10735/kresd
udp        0      0 *                           10735/kresd
udp        0      0 :::53                   :::*                                10735/kresd

shows me it’s listening only on 53 (though I wonder what the other ports are for).

If I tail -f messages and restart kresd I see only this appear:

2021-03-10 21:19:25 err kresd[12059]: deprecation WARNING: use --noninteractive instead of --forks=1
2021-03-10 21:19:25 info kresd[12059]: [system] warning: hard limit for number of file-descriptors is only 4096 but recommended value is 524288
2021-03-10 21:19:25 info kresd[12059]: [ta_update] refreshing TA for .
2021-03-10 21:19:25 info kresd[12059]: [ta_update] next refresh for . in 23.9075 hours
2021-03-10 21:19:26 warning /[]: Add_lease, hostname check failed
2021-03-10 21:19:26 warning /[13938]: Last message 'Add_lease, hostname ' repeated 4 times, suppressed by syslog-ng on Cerberus.lan
2021-03-10 21:19:26 info /[]: Refresh kresd leases

which highlights to config errors (that I have no idea how to fix, a --noninteractive error and hard limit for number of file descriptors warning), but does not show any other issues, neither a sign of success or failure in loading the http module.

What am I doing wrong?

  • dependencies are still missing: Knot - enable HTTP
  • just loading the module wouldn’t turn this feature on
  • for most people the verbose logs aren’t readable enough to make them understand the cause of the problem (by our experience that’s so, so typically send us the logs)
  • the standard Turris way to gather logs is described in wiki: