[arch-general] Ping: 100% package loss

Sun Jan 29 13:25:02 UTC 2017

Thorsten Jolitz via arch-general <arch-general at archlinux.org> writes:

Hello,
following up to my own post again.

> Thorsten Jolitz via arch-general <arch-general at archlinux.org> writes:
>
> Hello,
> following up to my own post:
>
>> Marcel Hoppe via arch-general <arch-general at archlinux.org> writes:
>>
>> Hi Marcel, Hi Robin,
>>
>> thanks for your answers.
>>
>>> I resolved the same problem 😉 on my systems I run the
>>> networkmanager and
>>> this works long time - I'm not sure but I think the problem was that
>>> systemd gets or starts its own revolver service. After disabling it and
>>> deleting the linked resolve.conf the networkmanager creates it after a
>>> restart again and it works.
>>
>> In Arch it's all about IPv6 it seems, so I did not want to disable
>> it. Instead I deleted the old resolve.conf (it was replaced
>> automatically by systemd). I replaced one (probably wrong) reference to
>> IPv4 with IPv6 in my config, and started and enabled the network service
>> again.
>>
>> Now pinging works for both protocols, and my internet connection went
>> from incredibly slow to more or less acceptable.
>
> So yesterday it worked, pinging IPv6 adresses as well as having
> acceptable internet connection speed.
> Today not anymore, although I did not change anything (ok, wasn't there
> a kernel update yesterday?)
>
> Any help interpreting this messages would be appreciated:
> ,----
> | Jan 28 13:03:58 arch systemd[1]: Timed out waiting for device
> | sys-subsystem-net-devices-service.device.
> | -- Subject: Unit sys-subsystem-net-devices-service.device has failed
> | -- Defined-By: systemd
> | -- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
> | --
> | -- Unit sys-subsystem-net-devices-service.device has failed.
> | --
> | -- The result is timeout.
> | Jan 28 13:03:58 arch systemd[1]: Dependency failed for dhcpcd on
> | service.
> | -- Subject: Unit dhcpcd at service.service has failed
> | -- Defined-By: systemd
> | -- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
> | --
> | -- Unit dhcpcd at service.service has failed.
> | --
> | -- The result is dependency.
> | Jan 28 13:03:58 arch systemd[1]: dhcpcd at service.service: Job
> | dhcpcd at service.service/start failed with result 'dependency'.
> `----

This issue seems to be related to versions of systemd > 230, and I followed
the advice in the wiki (adding a custom 'systemd-user-sessions.service'
to '/etc/systemd/system/multi-user.target.wants'), and the symptom seems
to have gone away.

But not so the IPv6 and internet speed problems. I went to all kinds of
wiki articles about network configuration, IPv6 and DHCP, and adapted my
config in several places to advices from that pages, but to no avail.

Now I don't even know how to further diagnose the problem, being unable
to ping IPv6 addresses and incredibly slow internet connections seem to
be the only visible symptom, everything else looks just fine.

And the most surprising thing is, that it worked for one single moment,
see the PS, and stopped working after the next reboot - with all what I
tried to make it work still untouched and in place.

Any further tipps here?
TIA

> Just time out because of lousy internet connecton?
>
> ,----
> |   /run:
> |   -rw-r--r--  1 root root    5 28. Jan 09:56 dhcpcd.pid
> `----
> ,----
> | $ ps 1952
> |   PID TTY      STAT   TIME COMMAND
> |  1952 ?        Ss     0:00 dhcpcd
> `----
>
> At startup, when calling dhcpcd, messages indicate that things worked
> out, but it seems the service never really started?
>
>> PS
>> ,----
>> | $ ping -c 3 www.google.com
>> | PING www.google.com(waw02s07-in-x04.1e100.net
>> | (2a00:1450:401b:802::2004%2)) 56 data bytes
>> | 64 bytes from waw02s07-in-x04.1e100.net (2a00:1450:401b:802::2004):
>> | icmp_seq=1 ttl=53 time=54.2 ms
>> | 64 bytes from waw02s07-in-x04.1e100.net (2a00:1450:401b:802::2004):
>> | icmp_seq=2 ttl=53 time=49.2 ms
>> | 64 bytes from waw02s07-in-x04.1e100.net (2a00:1450:401b:802::2004):
>> | icmp_seq=3 ttl=53 time=50.2 ms
>> | 
>> | --- www.google.com ping statistics ---
>> | 3 packets transmitted, 3 received, 0% packet loss, time 2003ms
>> | rtt min/avg/max/mdev = 49.253/51.266/54.294/2.187 ms
>> `----
>>
>> ,----
>> | $ ping -c 3 ipv6.google.com
>> | PING ipv6.google.com(waw02s08-in-x0e.1e100.net
>> | (2a00:1450:401b:803::200e%2)) 56 data bytes
>> | 64 bytes from waw02s08-in-x0e.1e100.net (2a00:1450:401b:803::200e):
>> | icmp_seq=1 ttl=53 time=82.3 ms
>> | 64 bytes from waw02s08-in-x0e.1e100.net (2a00:1450:401b:803::200e):
>> | icmp_seq=2 ttl=53 time=48.7 ms
>> | 64 bytes from waw02s08-in-x0e.1e100.net (2a00:1450:401b:803::200e):
>> | icmp_seq=3 ttl=53 time=48.8 ms
>> | 
>> | --- ipv6.google.com ping statistics ---
>> | 3 packets transmitted, 3 received, 0% packet loss, time 2003ms
>> | rtt min/avg/max/mdev = 48.754/59.986/82.355/15.819 ms
>> `----
>>
>> ,----
>> | $ ping -c 3 www.web.de
>> | PING www.g-ha-web.de (82.165.230.17) 56(84) bytes of data.
>> | 64 bytes from bap.web.de (82.165.230.17): icmp_seq=1 ttl=247
>> | time=226 ms
>> | 64 bytes from bap.web.de (82.165.230.17): icmp_seq=2 ttl=247 time=27.8
>> | ms
>> | 64 bytes from bap.web.de (82.165.230.17): icmp_seq=3 ttl=247 time=27.4
>> | ms
>> | 
>> | --- www.g-ha-web.de ping statistics ---
>> | 3 packets transmitted, 3 received, 0% packet loss, time 2003ms
>> | rtt min/avg/max/mdev = 27.429/93.814/226.124/93.557 ms
>> `----
>>
>>> Greets,
>>> Marcel
>>>
>>> Am 22.01.2017 17:31 schrieb "Robin via arch-general" <
>>> arch-general at archlinux.org>:
>>>
>>>> Hi,
>>>>
>>>> seems like a ipv6 related problem. Your DNS Lookup is v4, first ping is
>>>> v6 and second ping is v4 again. Try to ping the ipv4 address of
>>>> google.com. But im pretty sure this is not arch related.
>>>>
>>>> Maybe you should try to disable ipv6 system wide and then check if it
>>>> works.
>>>>
>>>> Cheers,
>>>> Robin
>>>> > Hello List,
>>>> > sometime age my msmtp imap connections just stopped working.
>>>> > Investigating the cause, I checked quite a lot of things, and came
>>>> > across the ping "100% package loss" problem:
>>>>
>>>

-- 
cheers,
Thorsten