Greetings.
Something changed in our work network configuration or my Linux laptop around Christmas. I now have serious network issues.
The following happens all the time.
---------------------------------------------------------------------- 1. I am unable to contact a server
http://en.wikipedia.org/ Browser says ERR_ADDRESS_UNREACHABLE
2. Name server works just fine
[jarmo@localhost tmp]$ dig en.wikipedia.org +short 91.198.174.192
3. traceroute shows issues
[jarmo@localhost tmp]$ traceroute en.wikipedia.org traceroute to en.wikipedia.org (91.198.174.192), 30 hops max, 60 byte packets 1 localhost.localdomain (172.16.43.148) 3004.217 ms !H 3004.195 ms !H 3004.187 ms !H
4. The problem is solved temporarily by flushing the route cache
[jarmo@localhost tmp]$ ip ro get 91.198.174.192 91.198.174.192 via 91.198.174.192 dev em1 src 172.16.43.148 cache <redirected> [jarmo@localhost tmp]$ su -c "ip route flush cache" [jarmo@localhost tmp]$ ip ro get 91.198.174.192 91.198.174.192 via 172.16.40.1 dev em1 src 172.16.43.148 cache
Now the address works ok for a while, but then I get connection problems again quite quickly. ----------------------------------------------------------------------
I have no idea whether the problem is in my laptop or in the network. I have no problems in other networks. On the other hand, Windows machine in this network work without issues.
Any ideas? I am running an up-to-date Fedora 20 with
[jarmo@localhost tmp]$ uname -a Linux localhost.localdomain 3.17.8-200.fc20.x86_64 #1 SMP Thu Jan 8 23:26:57 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
[jarmo@localhost tmp]$ yum info iproute Loaded plugins: langpacks, refresh-packagekit Installed Packages Name : iproute Arch : x86_64 Version : 3.14.0 Release : 2.fc20
Jarmo
On 02/02/2015 12:09 PM, Jarmo Hurri wrote:
Greetings.
Something changed in our work network configuration or my Linux laptop around Christmas. I now have serious network issues.
The following happens all the time.
- I am unable to contact a server
http://en.wikipedia.org/ Browser says ERR_ADDRESS_UNREACHABLE
- Name server works just fine
[jarmo@localhost tmp]$ dig en.wikipedia.org +short 91.198.174.192
- traceroute shows issues
[jarmo@localhost tmp]$ traceroute en.wikipedia.org traceroute to en.wikipedia.org (91.198.174.192), 30 hops max, 60 byte packets 1 localhost.localdomain (172.16.43.148) 3004.217 ms !H 3004.195 ms !H 3004.187 ms !H
- The problem is solved temporarily by flushing the route cache
[jarmo@localhost tmp]$ ip ro get 91.198.174.192 91.198.174.192 via 91.198.174.192 dev em1 src 172.16.43.148 cache <redirected> [jarmo@localhost tmp]$ su -c "ip route flush cache" [jarmo@localhost tmp]$ ip ro get 91.198.174.192 91.198.174.192 via 172.16.40.1 dev em1 src 172.16.43.148 cache
Now the address works ok for a while, but then I get connection problems again quite quickly.
I have no idea whether the problem is in my laptop or in the network. I have no problems in other networks. On the other hand, Windows machine in this network work without issues.
Any ideas? I am running an up-to-date Fedora 20 with
[jarmo@localhost tmp]$ uname -a Linux localhost.localdomain 3.17.8-200.fc20.x86_64 #1 SMP Thu Jan 8 23:26:57 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
[jarmo@localhost tmp]$ yum info iproute Loaded plugins: langpacks, refresh-packagekit Installed Packages Name : iproute Arch : x86_64 Version : 3.14.0 Release : 2.fc20
Jarmo
Duplicate IP Addresses somewhere in your network?
suomi
fedora fedora@ayni.com writes:
On 02/02/2015 12:09 PM, Jarmo Hurri wrote:
Duplicate IP Addresses somewhere in your network?
Might be the case. Can Windows handle this situation? In order to contact my network admin, I need something to explain why it's only my laptop that has issues.
Jarmo
On 02/02/15 19:09, Jarmo Hurri wrote:
Greetings.
Something changed in our work network configuration or my Linux laptop around Christmas. I now have serious network issues.
The following happens all the time.
- I am unable to contact a server
http://en.wikipedia.org/ Browser says ERR_ADDRESS_UNREACHABLE
- Name server works just fine
[jarmo@localhost tmp]$ dig en.wikipedia.org +short 91.198.174.192
- traceroute shows issues
[jarmo@localhost tmp]$ traceroute en.wikipedia.org traceroute to en.wikipedia.org (91.198.174.192), 30 hops max, 60 byte packets 1 localhost.localdomain (172.16.43.148) 3004.217 ms !H 3004.195 ms !H 3004.187 ms !H
- The problem is solved temporarily by flushing the route cache
[jarmo@localhost tmp]$ ip ro get 91.198.174.192 91.198.174.192 via 91.198.174.192 dev em1 src 172.16.43.148 cache <redirected> [jarmo@localhost tmp]$ su -c "ip route flush cache" [jarmo@localhost tmp]$ ip ro get 91.198.174.192 91.198.174.192 via 172.16.40.1 dev em1 src 172.16.43.148 cache
Now the address works ok for a while, but then I get connection problems again quite quickly.
I have no idea whether the problem is in my laptop or in the network. I have no problems in other networks. On the other hand, Windows machine in this network work without issues.
Any ideas? I am running an up-to-date Fedora 20 with
After you clear the cache can you try
echo 0 > /proc/sys/net/ipv4/conf/all/accept_redirects
and make sure you get
cat /proc/sys/net/ipv4/conf/all/accept_source_route 0
Then see if the problem returns.
Duplicate IP Addresses somewhere in your network?
Jarmo Hurri:
Might be the case. Can Windows handle this situation?
I'm not aware of any network thingummy that will nicely handle another device using the same IP. They're supposed to be unique on a network, and I'd go as far as saying that I'm sure they absolutely need to be.
Ed Greshko ed.greshko@greshko.com writes:
echo 0 > /proc/sys/net/ipv4/conf/all/accept_redirects
and make sure you get
cat /proc/sys/net/ipv4/conf/all/accept_source_route 0
That didn't help, because for some (weird?) reason after setting accept_redirects=0 for all, my system still had accept_redirects=1 for em1. However, setting accept_redirects=0 for em1 seems to have helped, at least in a short test. So thanks a bunch!
Can you explain what is happening here? I am assuming that Windows has this configured suitably in the first place, and that is why I am seeing the issue in Linux only.
Jarmo
On 02/03/15 17:56, Jarmo Hurri wrote:
Ed Greshko ed.greshko@greshko.com writes:
echo 0 > /proc/sys/net/ipv4/conf/all/accept_redirects
and make sure you get
cat /proc/sys/net/ipv4/conf/all/accept_source_route 0
That didn't help, because for some (weird?) reason after setting accept_redirects=0 for all, my system still had accept_redirects=1 for em1. However, setting accept_redirects=0 for em1 seems to have helped, at least in a short test. So thanks a bunch!
An IP redirect is sent by a (typically) router when it determines that a packet it has received would be better routed by an alternate path and it informs the source by sending it an ICMP redirect. In this case, it would seem your system received a redirect telling your system that 91.198.174.192 is directly connected to em1.
I would use wireshark to try and trap ICMP messages looking for the source of the redirects.
Can you explain what is happening here? I am assuming that Windows has this configured suitably in the first place, and that is why I am seeing the issue in Linux only.
I don't know anything about Windows and what they do. Sorry.
-
Ed Greshko -
fedora -
Jarmo Hurri -
Tim