Public bug reported:

In Description: Ubuntu 18.04.1 LTS
Release:        18.04

Occasionally an attempt to open a connection fails, or an attempt to
transfer a file, or do something else that works by opening a socket
fails.  Other things that you would expect to proceed normally with an
already opened socket continue to do so. That's the case with TCP, at
least.  One might draw a conclusion based on the log contents described
below that it would probably not so be for UDP, but I haven't tried to
verify that.

Examining syslog reveals groupings of strange repetitive behavior as
demonstrated in the 3 successive groups copied from syslog below.

I'm not certain of exactly when this behavior first appeared. I don't
know if it's related, but since it appeared, I have seen no more
instances of a different  anomalous interface network interface behavior
(namely:  the network being unavailable upon system resumption following
system suspension, and requiring performance of at least once more
system suspension in order to restore network connectivity)

The following are the most concise extracts I could make from the logs to 
demonstrate the issue. See the attachment for more:
Oct 31 00:07:57 notebook ntpd[5122]: Deleting interface #124 enp2s0, 
192.168.0.215#123, interface stats: received=72, sent=72, dropped=0, 
active_time=211 secs
Oct 31 00:07:57 notebook ntpd[5122]: Deleting interface #125 enp2s0, 
fe80::540b:bc14:dd5f:304f%2#123, interface stats: received=0, sent=0, 
dropped=0, active_time=211 secs
Oct 31 00:07:57 notebook kernel: [16414.973867] atl1 0000:02:00.0: enp2s0 link 
is up 100 Mbps full duplex
Oct 31 00:07:57 notebook NetworkManager[975]: <info>  [1540969677.5690] device 
(enp2s0): carrier: link connected
Oct 31 00:07:59 notebook ntpd[5122]: Listen normally on 126 enp2s0 
192.168.0.215:123
Oct 31 00:07:59 notebook ntpd[5122]: Listen normally on 127 enp2s0 
[fe80::540b:bc14:dd5f:304f%2]:123
Oct 31 00:12:42 notebook kernel: [16699.579883] atl1 0000:02:00.0: enp2s0 link 
is down

Oct 31 00:12:43 notebook ntpd[5122]: Deleting interface #126 enp2s0, 
192.168.0.215#123, interface stats: received=87, sent=88, dropped=0, 
active_time=284 secs
Oct 31 00:12:43 notebook ntpd[5122]: Deleting interface #127 enp2s0, 
fe80::540b:bc14:dd5f:304f%2#123, interface stats: received=0, sent=0, 
dropped=0, active_time=284 secs
Oct 31 00:12:43 notebook kernel: [16701.175235] atl1 0000:02:00.0: enp2s0 link 
is up 100 Mbps full duplex
Oct 31 00:12:43 notebook NetworkManager[975]: <info>  [1540969963.7701] device 
(enp2s0): carrier: link connected
Oct 31 00:12:45 notebook ntpd[5122]: Listen normally on 128 enp2s0 
192.168.0.215:123
Oct 31 00:12:45 notebook ntpd[5122]: Listen normally on 129 enp2s0 
[fe80::540b:bc14:dd5f:304f%2]:123
Oct 31 00:13:09 notebook kernel: [16726.743280] atl1 0000:02:00.0: enp2s0 link 
is down

Oct 31 00:13:10 notebook ntpd[5122]: Deleting interface #128 enp2s0, 
192.168.0.215#123, interface stats: received=20, sent=20, dropped=0, 
active_time=25 secs
Oct 31 00:13:10 notebook ntpd[5122]: Deleting interface #129 enp2s0, 
fe80::540b:bc14:dd5f:304f%2#123, interface stats: received=0, sent=0, 
dropped=0, active_time=25 secs
Oct 31 00:13:10 notebook NetworkManager[975]: <info>  [1540969990.9253] device 
(enp2s0): carrier: link connected
Oct 31 00:13:10 notebook kernel: [16728.330285] atl1 0000:02:00.0: enp2s0 link 
is up 100 Mbps full duplex
Oct 31 00:13:12 notebook ntpd[5122]: Listen normally on 130 enp2s0 
192.168.0.215:123
Oct 31 00:13:12 notebook ntpd[5122]: Listen normally on 131 enp2s0 
[fe80::540b:bc14:dd5f:304f%2]:123
Oct 31 00:16:30 notebook kernel: [16927.574126] atl1 0000:02:00.0: enp2s0 link 
is down


While not scientific, I suspect the following  view of the (brief enough to 
permit reliable TCP byte streams, as verified by md5sum) timing between "link 
down" and "link up" events (as opposed to the much longer amount of time 
elapsing between "link up"  and the next down event) might help to explain why 
the problem is not more symptomatic than it is:

[Oct30 20:23] atl1 0000:02:00.0: enp2s0 link is down
[  +1.544088] atl1 0000:02:00.0: enp2s0 link is up 100 Mbps full duplex
[Oct30 20:24] atl1 0000:02:00.0: enp2s0 link is down
[  +1.630772] atl1 0000:02:00.0: enp2s0 link is up 100 Mbps full duplex
[Oct30 20:32] atl1 0000:02:00.0: enp2s0 link is down
[Oct30 20:33] atl1 0000:02:00.0: enp2s0 link is up 100 Mbps full duplex
[Oct30 20:38] atl1 0000:02:00.0: enp2s0 link is down
[  +1.626526] atl1 0000:02:00.0: enp2s0 link is up 100 Mbps full duplex
[Oct30 20:42] atl1 0000:02:00.0: enp2s0 link is down
[  +1.578927] atl1 0000:02:00.0: enp2s0 link is up 100 Mbps full duplex

As before, this is brief, but enough to be illustrative, and  more can
be found in the tarred log info.

** Affects: linux (Ubuntu)
     Importance: Undecided
         Status: Incomplete


** Tags: bionic ethernet frequent-seesawing interface link-down link-up 
seesawing

** Attachment added: "dmesg,loggreps.tar"
   
https://bugs.launchpad.net/bugs/1800888/+attachment/5207589/+files/dmesg%2Cloggreps.tar

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1800888

Title:
  interfaces repetitively deleted in apparent  synchrony with up/down
  network interface seesawing

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1800888/+subscriptions

-- 
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to