Tag Info

Hot answers tagged

5

There's no difference; dev is optional provided the device name can't be mistaken for another keyword that the ip command understands. For instance, if you had an interface named bridge then you would have to use dev as bridge is also a keyword that ip uses. This is undocumented behavior, but it certainly appears to work. You'll know if you ran afoul of ...


5

That image says that the maximum UDP transfer rate over 1gbps ethernet is 119635891 Bytes per second. Multiple 119635891 with 8 (one byte is 8 bits) and you get 957087128 == 957,08mbps


3

After checking that local interface is up (ifup lo0 or similar). Check for routes, firewall or iptables problem. Here it was a problem from the file located here /proc/sys/net/ipv4/icmp_echo_ignore_all When set to 1 : There is no response to ICMP Requests, so ping will not work. When set to 0 : Replies to ICMP Requests are enable. Check if sets to 1. If ...


2

/sbin/route add net 136.170.192.0 netmask 255.255.252.0 gw 10.240.100.110 Try route add -net .... I suspect that error message comes when your embedded device tries to resolve the name net to an address or network range.


2

With your current configuration your default gateway is the tunnel itself (do1). So when you ping 104.000.65.38 from the internet the packets arrive to eth0 but they return via the do1 tunnel instead. You will need to do connection marking with iptables and set up a secondary routing table with default gateway 104.000.64.1 so that packets that arrive to ...


2

Found it! This command is not enough: ifconfig eth0 up Instead I must do: ifconfig eth0 up sysctl -w net.ipv6.conf.eth0.disable_ipv6=0 Only the latter does restore the IPv6 link local address on eth0. In other words, my distro seems to have some code hidden somewhere which dynamically changes net.ipv6.conf.eth0.disable_ipv6 every time it is asked to ...


1

As I don't see it in your shown documentation, I'm going to assume you forgot to enable IP forwarding. Run the following command to verify that this is indeed the case: sysctl net.ipv4.ip_forward If IP forwarding is indeed disabled you should see: net.ipv4.ip_forward = 0 If this is your problem, run: sysctl -w net.ipv4.ip_forward=1 Possibly your ...


1

As I understood you right, you trying to create 70000 child classes like that (1:1 - 1:70000), am I right? if I am right, have you tried to add some intermediate classes (for example 1:1 - 1:9999, 2:1 -2:9999, etc.) ? You may need to create more complex tree... For example you could classify traffic by subnets at first and by ip at the leaf classes.


1

What I see in your packet trace is congestion control reacting to packet loss. The client starts out sending an initial 9 segments followed by slow start, where it send two more segments each time it receives an ACK packet. The slow start algorithm continues until the first duplicate ACK from the server indicates that a packet has been lost. This happens ...


1

In the official OpenVPN documentation you can find: [...] --route-nopull When used with --client or --pull, accept options pushed by server EXCEPT for routes and dhcp options like DNS servers. When used on the client, this option effectively bars the server from adding routes to the client's routing table, however note that this option still ...


1

I had to do exactly this. I found that the answer suggested by @Thusitha was the correct way to do it for new kernels. Tested with the Debian wheezy kernel 3.2.0-4 and iproute (from where the tc command comes from) version 20120521-3+b3 Here is the complete script, the tc filter lines being almost exactly as specified by @Thusitha function qos() { ...


1

Your clients on the other end need the routing information as well, so I believe you'll need to have a gateway set for them to talk the other way around unless your VPN server currently acts as the default gateway on the 192.168.178.x network. (if your firewall is set up right), the machines on the local LAN would need: route add -net 10.0.0.0 netmask ...


1

From RFC 1812 : (e) { 127, } Internal host loopback address. Addresses of this form MUST NOT appear outside a host. The kernel uses net.ipv4.conf.<interface>.route_localnet to make 127.0.0.0/8 routable or not (actually consider source/destination IP addresses of this kind as martian packets or not). I completely disencourage you to ...


1

Linux has a weak host model and the network stack doesn't really care which interface an TCP/IP packet arrives on. If it is for a configured IP-address it will be accepted. That does not take into account ARP, firewall rules and routing.



Only top voted, non community-wiki answers of a minimum length are eligible