I have a relatives PC running Suse 13.1. It has been working fine for a long time just fine, but recently their Ethernet has been taking a substantially long time to come up after a resume from suspend. My interpretation of the logs is that it isn’t properly renewing its connections, but reading the man pages of NM and dhcpcd didn’t produce any results for me. Here is their NM log:
There are two resumes here - one after 5 hours, in which case it takes 6 seconds to renew the connection, and then one where I just toggle on / off networking and the connection comes back in a second.
I’ll also include a log from one of my machines on their network, to just show that NM with, in this case dhclient, can renew properly even after upwards of 12 hours of downtime, so I don’t think its an issue with their router.
Here are the gaps, first two seconds between disconnected and bringing the carrier up:
2014-05-02T08:34:35.441001-04:00 RubyTempest NetworkManager: <info> NetworkManager state is now DISCONNECTED
2014-05-02T08:34:37.592211-04:00 RubyTempest NetworkManager: <info> (enp3s0): carrier now ON (device state 20)
Then three seconds in preinit still, this time with dhclient:
This seems to indicate it is some configuration in networkmanager itself that is making it bring the network back up so slowly. I’m curious why it isn’t doing instant renewals, is there a tweak in something other than NetworkManager.conf that lets you expend the dhcp renewal period?