Due to a lack of failures it has been a while since I last wrote, but I think I may have found a possible cause of the problem. Confirmation from other sources would be necessary, of course.
For the previous participants, and anyone else who lands here, I think that the failure may go like this:- - a network problem prevents reception of a D.H.C.P. lease offer from the server - NetworkManager starts a new D.H.C.P. client thread, but the old thread is still active and the lease has not expired - these two steps may be repeated several times, starting several D.H.C.P. threads, with identical or different addresses, possibly receiving and transmitting traffic - some time (possibly hours) later, another instance of the client has leased the address, the original client's lease expires - avahi surrenders the address; the host has lost its network connection In my case, some months earlier, it became apparent that the D.H.C.P. server was connected to a network switch port that was occasionally dropping packets - hence errors were reported for the server's and client's network adaptors. The packet retry count and network load increased. I changed the physical network arrangement and there were no more failures - until I switched off the server for maintenance. Here is a test which might support this theory. Note that the server was not available between 05:38 and 07:17 on 2018-01-24. The client's network connection was not lost until 2018-01-30. [LinuxMint] graham@yoyo $ ps -ef | grep -v "grep" | grep -e "PID" -e "dhc" UID PID PPID C STIME TTY TIME CMD root 2444 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 3327 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 4550 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 5070 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 6094 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 6616 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 7605 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 8179 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 9152 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 9442 1 0 Jan24 ? 00:00:00 /sbin/dhclient -v eth0 root 21279 11291 0 03:26 ? 00:00:00 /sbin/dhclient -d -q -sf /usr/lib/NetworkManager/nm-dhcp-helper -pf /var/run/dhclient-eth0.pid -lf /var/lib/NetworkManager/dhclient-3339a4b3-ef0a-3db1-acf1-9b810d040dfe-eth0.lease -cf /var/lib/NetworkManager/dhclient-eth0.conf eth0 Tue Jan 30 08:38:33 ~ The start times of those dhclient instances were:- 05:50 05:55 06:05 06:11 06:21 06:27 06:37 06:43 06:53 06:59. The log server had many instances of exchanges between kevin the server and yoyo the client like this:- 2018-01-30T07:25:11+00:00 kevin dhcpd[3612]: DHCPREQUEST for 192.168.42.187 from 00:25:64:72:72:14 via enp0s19 2018-01-30T07:25:11+00:00 kevin dhcpd[3612]: DHCPACK on 192.168.42.187 to 00:25:64:72:72:14 via enp0s19 2018-01-30T07:25:17+00:00 yoyo dhclient[2444]: DHCPREQUEST of 192.168.42.161 on eth0 to 192.168.42.129 port 67 (xid=0x182b3059) 2018-01-30T07:25:17+00:00 kevin dhcpd[3612]: reuse_lease: lease age 306 (secs) under 25% threshold, reply with unaltered, existing lease for 192.168.42.161 2018-01-30T07:25:17+00:00 kevin dhcpd[3612]: DHCPREQUEST for 192.168.42.161 from 00:25:64:72:72:14 via enp0s19 2018-01-30T07:25:17+00:00 kevin dhcpd[3612]: DHCPACK on 192.168.42.161 to 00:25:64:72:72:14 via enp0s19 2018-01-30T07:25:18+00:00 yoyo dhclient[8179]: DHCPREQUEST of 192.168.42.162 on eth0 to 255.255.255.255 port 67 (xid=0x537943a) 2018-01-30T07:25:18+00:00 kevin dhcpd[3612]: reuse_lease: lease age 4397 (secs) under 25% threshold, reply with unaltered, existing lease for 192.168.42.162 I have tried to be brief in this report, but I can provide more information if anyone wants it. -- You received this bug notification because you are a member of Ubuntu Touch seeded packages, which is subscribed to network-manager in Ubuntu. https://bugs.launchpad.net/bugs/1586528 Title: Avahi-daemon withdraws address record Status in avahi package in Ubuntu: Confirmed Status in network-manager package in Ubuntu: Confirmed Bug description: For some reason, if I leave my Ubuntu VM up for a prolonged period of time the machine will lose connection to the network. ip addr shows that the nic port no longer has an address and an examination of the syslog shows this: May 27 14:19:38 matt-VirtualBox avahi-daemon[590]: Withdrawing address record for 10.0.2.15 on enp0s3. May 27 14:19:38 matt-VirtualBox avahi-daemon[590]: Leaving mDNS multicast group on interface enp0s3.IPv4 with address 10.0.2.15. May 27 14:19:38 matt-VirtualBox avahi-daemon[590]: Interface enp0s3.IPv4 no longer relevant for mDNS. for no known reason. The only reliable way to get the network to come back (that I have found) is a full reboot. ProblemType: Bug DistroRelease: Ubuntu 16.04 Package: avahi-daemon 0.6.32~rc+dfsg-1ubuntu2 ProcVersionSignature: Ubuntu 4.4.0-22.40-generic 4.4.8 Uname: Linux 4.4.0-22-generic x86_64 ApportVersion: 2.20.1-0ubuntu2 Architecture: amd64 CurrentDesktop: Unity Date: Fri May 27 15:11:34 2016 InstallationDate: Installed on 2015-10-22 (218 days ago) InstallationMedia: Ubuntu 15.10 "Wily Werewolf" - Release amd64 (20151021) SourcePackage: avahi UpgradeStatus: Upgraded to xenial on 2016-03-30 (58 days ago) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/avahi/+bug/1586528/+subscriptions -- Mailing list: https://launchpad.net/~touch-packages Post to : touch-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~touch-packages More help : https://help.launchpad.net/ListHelp
