Re: [Dnsmasq-discuss] ubus problem

2019-04-27 Thread e9hack
Am 10.04.2019 um 23:56 schrieb Simon Kelley: > If Harmut's config is unique, and works without openWRT talking to both > dnsmasq instances, then the solution may be just to turn off ubus on one > of the dnsmasq instances. I don't know what openwrt does with the dnsmasq data on ubus. I did turn

Re: [Dnsmasq-discuss] ubus problem

2019-04-11 Thread wkitty42
On 4/10/19 12:55 PM, Jan Willem Janssen wrote: There's one solution I can think of: making the name under which we register the UBus object configurable (with "dnsmasq" as default for backwards compatibility). It would allow multiple instances to be configured each with their own unique name.

Re: [Dnsmasq-discuss] ubus problem

2019-04-10 Thread Simon Kelley
On 10/04/2019 17:55, Jan Willem Janssen wrote: > On Mon, 2019-04-08 at 20:41 +0100, Simon Kelley wrote: >>> I've to give it some thought about how we could support multiple Dnsmasq >>> instances in >>> combination with UBus. Not sure how the DBus implementation would handle >>> this... >> >>

Re: [Dnsmasq-discuss] ubus problem

2019-04-10 Thread Jan Willem Janssen
On Mon, 2019-04-08 at 20:41 +0100, Simon Kelley wrote: > > I've to give it some thought about how we could support multiple Dnsmasq > > instances in > > combination with UBus. Not sure how the DBus implementation would handle > > this... > > It doesn't: the path is a compile-time parameter. >

Re: [Dnsmasq-discuss] ubus problem

2019-04-08 Thread wkitty42
On 4/8/19 3:58 PM, wkitt...@gmail.com wrote: is there some ID or signature that could be used to differentiate between separate dnsmasq instances? if so, one could specify that in the config and that could be used with dbus to separate the instances and how they communicate... i mixed up

Re: [Dnsmasq-discuss] ubus problem

2019-04-08 Thread wkitty42
On 4/8/19 1:52 PM, Jan Willem Janssen wrote: I've to give it some thought about how we could support multiple Dnsmasq instances in combination with UBus. Not sure how the DBus implementation would handle this... is there some ID or signature that could be used to differentiate between

Re: [Dnsmasq-discuss] ubus problem

2019-04-08 Thread Simon Kelley
> > I've to give it some thought about how we could support multiple Dnsmasq > instances in > combination with UBus. Not sure how the DBus implementation would handle > this... It doesn't: the path is a compile-time parameter. It's not clear that the entities on the other end of the UBus

Re: [Dnsmasq-discuss] ubus problem

2019-04-08 Thread Jan Willem Janssen
Hi, On Mon, 2019-04-08 at 16:24 +0200, e9hack wrote: > From the first I see: > daemon.info dnsmasq[1808]: Connected to system UBus > daemon.info dnsmasq[1808]: UBus support enabled: connected to system bus > > From the second I see: > daemon.err dnsmasq[1809]: Cannot add object to UBus: Invalid

Re: [Dnsmasq-discuss] ubus problem

2019-04-08 Thread Simon Kelley
@Jan? (I suspect that nothing has changed, except that a previously silent error is now no longer silent, but it would be nice to confirm this, and maybe explicitly consider this case.) Simon On 08/04/2019 15:24, e9hack wrote: > Hi, > > I'm using the latest dnsmasq version with openwrt.

[Dnsmasq-discuss] ubus problem

2019-04-08 Thread e9hack
Hi, I'm using the latest dnsmasq version with openwrt. There are two instances running. One provides dhcpv4, dhcpv6 and dns to several networks, the other one dhcpv4 to one network only. It looks like, that the second instance has a problem with ubus: Sun Apr 7 18:37:34 2019 daemon.info