Solved Looking for a network monitoring daemon

Hmm, I guess pool.ntp.br has a really low TTL, it ages out of the unbound cache very quickly, and that's why you're having the name resolution failures. Have you considered running a local ntpd(8)?
 
I do not consider the failing requests of the ntp daemon as an error, but as expected behaviour when the internet connection is down.

Clearly, the ISP had a problem on their side, and on my side various dhclient child processes tried to reestablish the connection for more than one hour, but finally the dhclient mother process was shutdown by the kernel with a SIGSEV, because it shot itself in the food (segmentation fault). So once the ISP had resolved the problem outside, my system could not recover because dhclient was already dead.

For the time being, I build & installed a dhclient binary with debug symbols compiled into (CFLAGS="-g -O0"), and I disabled the pingorrestart.sh script wich I mentioned in post #12. For now I have to wait until the next incident, and when the network is down the other day, I could debug the core dump like so: https://obsigna.com/articles/1566473244.html
 
You must log in or register to reply here.
Back
Top