Comment 0 for bug 956918

Revision history for this message
Didi (d10r) wrote :

The following issue occured several times over the last month with Ubuntu 11.10 x86_64, kernel 3.0.0-16-generic.
It happened from the beginning (3.0.0-15). I didn't have this Computer before, so I don't know if the issue is specific to this Ubuntu / kernel version.

Sometimes, after waking up from suspend, single remote hosts are not reachable. It may be the first host the system tries to connect to (or may be the last one connected to before susped), but I'm not sure about that.
"not reachable" means there's no route. It's not a DNS issue because trying to ping the IP itself also fails. Example ping output:

$ ping 80.237.133.28
PING 80.237.133.28 (80.237.133.28) 56(84) bytes of data.
From 78.104.151.116 icmp_seq=1 Destination Host Unreachable
...

at the same time:
$ ping google.com
PING google.com (209.85.148.139) 56(84) bytes of data.
64 bytes from fra07s07-in-f139.1e100.net (209.85.148.139): icmp_req=1 ttl=57 time=15.3 ms
...

here, 78.104.151.116 is my own IP in this network.
The destination IP is definitely reachable, I can reach it at the same moment on a different machine in the same network. Or from my own machine if I first ssh to another machine and connect from there.

The issue is hardly a network interface driver issue, because this already happened with both WiFi and Ethernet connection.

It looks like the issue doesn't happen in all networks. In my home network it didn't happen yet, but already in two other networks. I don't know what makes the difference.

As I said, the issue is isolated to single destination hosts. At the same time, the connection works normally for all other destinations.
I could solve this only by reboot up to now. Things I tried:
- restart networking
- suspend / wakeup again
- disconnect / reconnect network interface
- switch network interface (WiFi / Ethernet)
- ip route flush cache
- ip neigh flush all

I am no networking expert, but a sysadmin of the university network took a look at my Computer when the issue appeared. He couldn't find out the reason, but suspected it to be a kernel issue.

I think the posting at http://askubuntu.com/questions/100717/some-ips-unreachable-after-resume-from-suspend describes the same issue, but I didn't get any further feedback there.

For any further advice on how I could gather more useful information / try to fix it I would be very helpful.

$ uname -srm
Linux 3.0.0-16-generic x86_64