[SOLVED] DHCP Seems to Stop Working Every 5-10 Minutes

/dev/dog · 12-28-2018, 03:22 AM

The problem is that my connection to the internet keeps getting screwed up every 5-10 minutes. I stay connected to my router, and I can still ping 8.8.8.8, but I cannot go to any websites. The solution I have found is to run

Code:

$ sudo dhclient

and it works again for the next 5-10 minutes. That is the only basis I am using to claim this is a problem with my DHCP client stopping.

Output after executing the above command:

Code:

dog@lapdog:~$ sudo dhclient
\RTNETLINK answers: File exists
autofs: unrecognized service
Starting fetch-ldap-cert (via systemctl): fetch-ldap-cert.service.

Another source of information:

Code:

dog@lapdog:~$ cat /var/log/messages
Dec 28 03:35:01 lapdog debian-edu-update-netblock: making sure netblock is disabled
Dec 28 03:35:34 lapdog autofs-reload: reloading autofs from dhcp exit hook
Dec 28 03:40:01 lapdog debian-edu-update-netblock: making sure netblock is disabled
Dec 28 03:42:27 lapdog autofs-reload: reloading autofs from dhcp exit hook
Dec 28 03:45:01 lapdog debian-edu-update-netblock: making sure netblock is disabled
Dec 28 03:45:39 lapdog autofs-reload: reloading autofs from dhcp exit hook
Dec 28 03:50:01 lapdog debian-edu-update-netblock: making sure netblock is disabled
Dec 28 03:51:19 lapdog autofs-reload: reloading autofs from dhcp exit hook
Dec 28 03:55:01 lapdog debian-edu-update-netblock: making sure netblock is disabled
Dec 28 03:56:12 lapdog autofs-reload: reloading autofs from dhcp exit hook
Dec 28 04:00:01 lapdog debian-edu-update-netblock: making sure netblock is disabled

The message "autofs-reload: reloading autofs from dhcp exit hook" corresponds to each time I had to run my fix.

/dev/dog · 12-28-2018, 03:31 AM

Right after I created the thread, I realized that the line "debian-edu-update-netblock: making sure netblock is disabled" occurs at the first second of every 5 minutes. I checked my crontabs and I found this:

Code:

dog@lapdog:~$ cat /etc/cron.d/debian-edu-config 
# Only dry-run cfengine cron job until we know it is safe to run on an installed
# system [pere 2003-08-30]
# Disable until it no longer run in "installation" mode. [pere 2003-09-01]
#9 4	* * *	root	if [ -x /usr/bin/cfengine-debian-edu ]; then /usr/bin/cfengine-debian-edu -n; fi
#
# Check every 5 minutes if network blocking should be enabled or disabled
*/5 * * * *	root	if [ -x /usr/sbin/debian-edu-update-netblock ]; then /usr/sbin/debian-edu-update-netblock auto > /dev/null; fi

I commented it out and now I will wait to see if it drops again.

EDIT: Solution did not work, internet still keeps not working. But at least those messages aren't showing up anymore in /var/log/messages. As a side not, is that important to keep?

I guess I should have first checked if my dhcp client is actually not running when I get the problems. That will be the next step when it screws up again.

/dev/dog · 12-28-2018, 03:49 AM

Code:

dog@lapdog:~$ sudo service dhcp status
● dhcp.service
   Loaded: not-found (Reason: No such file or directory)
   Active: inactive (dead)

However, even after I do the fix I found and I am able to use the internet again, the above still appears.

Code:

dog@lapdog:~$ sudo ps -Awl | grep dhclient
1 S     0   643     1  0  80   0 -  6350 -      ?        00:00:00 dhclient

(There are more than one of the above output but I only included one, the only difference between all of them is the PID)

/dev/dog · 12-28-2018, 11:26 PM

I think I found a suitable fix. At first I tried putting in my crontabs

Code:

* * * * * sudo dhclient

But I would occasionally find myself with my internet being screwed up again in the last 5-15 seconds of every minute. So then I also added

Code:

* * * * * ( sleep 30 ; sudo dhclient)

Which was working for me, but it was such a hacky solution that I wasn't satisfied. So I added Google's nameservers (8.8.8.8, and 8.8.4.4) to my /etc/resolvconf/resolv.conf.d/tail and removed the cron jobs and now it's all working. No problems in the last hour.

I am still not sure what was causing the problems, or if my /etc/resolv.conf kept being emptied, and a call to dhclient would solve the problem by repopulating the file, but it's solved now.

Thanks for all your help