[mythtv-users] Problems having mythbackend running after boot

[Bill Meek]

On 01/05/2014 11:32 AM, Roger Siddons wrote:
> On Sun, 05 Jan 2014 14:18:07 -0000, Bill Meek <keemllib at gmail.com> wrote:
>
>   > On 01/05/2014 06:34 AM, Marcel Jira wrote:
>
> <snip>
>
>   >> I have very little knowledge about the cause of this error, but it may
>   >> be related to the network configuration of my mythtv backend. I entered
>   >> 127.0.0.1 to the IP4-address and ::1 to the IP6-address.
>   >>
>   >> There's an error in the log saying
>   >> [CODE]E CoreContext serverpool.cpp:411 (listen) Failed listening on TCP
>   >> [fe80::5246:5dff:fe50:5509%eth0]:6543 - Error 9: The address is not
>   >> available[/CODE]
>   >>
>
> <snip>
>
>   >
>   > I had a similar problem and "solved" it by adding a sleep 5 just before
>   > mythbackend
>   > was started (in my case that was done in an Upstart .conf file.) The
>   > problem was that
>   > Upstart didn't guarantee that *all* of the addresses on the interface
>   > were up. That
>   > may care for the 3rd post boot startup in your log, but the 1st 2 are a
>   > bit scarier
>   > (Failed to connect to (127.0.0.1:6543) and I've not seen that.
>   >
>   > Good luck.
>   >
>
> In the first 2 logs it is trying to run as a slave backend; the 3rd is ok
>
> You could also try disabling IPv6, as described in
> http://code.mythtv.org/trac/ticket/11030

The 3rd failed too, on line 143:

     C CoreContext main_helpers.cpp:681 (run_backend) Backend exiting, MainServer initialization error.

Oops, I missed the slave part, so the 1st 2 don't count. While disabling
IPv6 is a good test (and the proper answer is that the startup script/
configuration should make sure ALL addresses are available,) just sleeping
while the interface settles works and allows me to keep IPv6 on.

Upstart, for example, doesn't guarantee all addresses are up when testing
with:  net-device-up IFACE!=lo or even net-device-up IFACE=eth0

-- 
Bill