#22247: Cannot run multiple instances of n2n edge as a service - not supported
by
current implementation in OPENWRT.
--------------------------+--------------------------------
Reporter: spam064@… | Owner: developers
Type: defect | Status: new
Priority: normal | Milestone:
Component: packages | Version: Chaos Calmer 15.05
Keywords: n2n edge vpn |
--------------------------+--------------------------------
Hello,
== Problems description ==
there are 2 issues which unable to start multiple (more than one)
instances of n2n edge for 2 separate private networks on one router.
== 1. it is not possible to define tap name in openwrt n2n config file ==
which then would have to be propoerly handled by /etc/init.d/n2n then.
According to man page it should be possible to set up multiple tap devices
http://linux.die.net/man/8/edge
Options
-d <name>
sets the TAP device name as seen in ifconfig. Only available on
Linux.
'''With current implementation in OPENWRT setting the TAP device name is
not supported in/via n2n config file.'''
Although edge application released for OPENWRT CC does support this
option.
== 2. For some reason it is not possible to run multiple instances of
"edge" as a service. ==
Once support for problem (1) is implemented, I realised that for some
reason is not possible to run multiple instances of edge as a service.
To enable to run multiple instances of n2n edge to set up multiple tap
devices edge has to be run not as a service but as an regular process.
== Proposed tested and working solution on CC to address issue (1) and (2)
==
1. introduce additional parameter "tapname" in n2n config file
example:
uci show n2n
n2n.@edge[0]=edge
n2n.@edge[0].netmask='255.255.255.0'
n2n.@edge[0].supernode='yoursupernode.com'
n2n.@edge[0].port='86'
n2n.@edge[0].community='community1'
n2n.@edge[0].key='yourkey1'
n2n.@edge[0].route='1'
n2n.@edge[0].httptunnel='1'
n2n.@edge[0].ipaddr='192.168.12.1'
'''n2n.@edge[0].tapname='edge0''''
n2n.@edge[1]=edge
n2n.@edge[1].netmask='255.255.255.0'
n2n.@edge[1].supernode='yoursupernode.com'
n2n.@edge[1].port='86'
n2n.@edge[1].community='yourcommunity2'
n2n.@edge[1].key='yourkey2'
n2n.@edge[1].route='1'
n2n.@edge[1].httptunnel='1'
n2n.@edge[1].ipaddr='10.10.10.1'
'''n2n.@edge[1].tapname='edge1''''
n2n.@supernode[0]=supernode
n2n.@supernode[0].port='86'
2. change the /etc/init.d/n2n the way to
a. support additional parameter "tapname" in n2n config file
b. change execution of edge as a process not a service to enable
multiple instances it on 1 host.
Here is a source of start_instance() function of '''/etc/init.d/n2n'''
with bolded proposed changes:
#!/bin/sh /etc/rc.common
START=90
start_instance() {
local cfg="$1"
config_get type "$cfg" TYPE
case "$type" in
edge)
config_get ipaddr "$cfg" 'ipaddr'
[ -n "$ipaddr" ] || return 1
config_get netmask "$cfg" 'netmask' '255.255.255.0'
config_get supernode "$cfg" 'supernode'
config_get port "$cfg" 'port'
config_get community "$cfg" 'community'
config_get key "$cfg" 'key'
config_get_bool route "$cfg" 'route' '0'
>'''config_get tapname "$cfg" 'tapname' 'edge0''''
[ "$route" = "1" ] && args='-r'
>'''#service_start /usr/sbin/edge $args -d $tapname -a
$ipaddr -s $netmask -c $community -k $key -l ${supernode}:${port}
>/usr/sbin/edge $args -d $tapname -a $ipaddr -s $netmask
-c $community -k $key -l ${supernode}:${port} &'''
;;
supernode)
[ -e /usr/sbin/supernode ] || return 1
config_get port "$cfg" port
[ -n "$port" ] || return 1
service_start /usr/sbin/supernode -l $port
;;
esac
}
[CUT]
'''Above change has been successfuly tested on CC release.'''
Thanks.
--
Ticket URL: <https://dev.openwrt.org/ticket/22247>
OpenWrt <http://openwrt.org>
Opensource Wireless Router Technology
_______________________________________________
openwrt-tickets mailing list
[email protected]
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-tickets
