I've searched this forum for other examples of this, and none of the other occurrances match my circumstances.

I've got R5E50 (I think) that's been running happily for a long while. When I had to move to schedulesdirect, I followed the instructions posted on this forum to rebuild mythtv and plugins from source and install. This fixed the schedules direct problem, plus a small additional patch got HDTV working for the first time ever! I never realized that HDTV has been broken in the code for so long -- I just figured I didn't have enough signal strength.

Anyway, the system has been running happily with the new code for a while (well, except my second NTSC tuner won't work anymore).

My master backend runs on a box (with a static IP), and my frontend in my entertainment center runs on a second box (with another static IP).

The power supply died in the master the other night (during an electrical storm) and after re-booting (several times), the frontends (either of them, master box and slave box) AND mythweb cannot connect to the backend.

Network interfaces are up just fine, I can SSH into either box just fine. The "lo" interface is up. mysql is up, and the DB is not corrupted. When I run setup, the IP addresses are still in there.

When I run "netstat -lnp | grep mythb" to see if mythbackend has port 6543 (or whatever it is) open, I see nothing. mythbackend is running successfully, and the log file doesn't say anything particularly interesting.

Help!

1. why isn't mythbackend listening on it's port?
2. How can I debug this?

I'm going to try to find where the port is opened in the source, but it's a lot of C++ and while I can struggle through it, it'll be a bear to find where that is in all this code.

I don't really want to re-install. That's such a windows-ish thing to do, and besides, I'm running the new version built from source, so the first thing I'd have to do to the new install is pi-- all over it. Plus, I want to keep all my recorded shows.

Thanks
Bill Dudley
New Jersey, USA

What makes you say "mythbackend is running successfully" when nothing can connect to it? What is the output of ps -ef | grep myth, ps -ef | grep -i sql and netstat -an | grep 6543 (6543 being the default mythtv port)?

When you say the "lo" interface is up I presume you mean on the backend box? Does the /etc/network/interfaces file correspond with the output from ifconfig _and_ the addresses you have defined "When you run setup"?

Did you change anything besides the power supply?

Thanks for the reply, "slowtolearn"

When I say "mythbackend is running successfully" I mean that I can see it with ps:

ps ax | grep mythb
8469 ? Ssl 0:00 /usr/bin/mythbackend --daemon -v none --logfile /var/log/mythtv/mythbackend.log --pidfile /var/run/mythtv/mythbackend.pid

netstat -an | grep 6543

returns null just like

netstat -lnp | grep mythb

does.

lo is up on both boxes. (I mentioned this since one of the other people to have this problem had a problem with lo not up.)

I did nothing besides change the power supply, and rebooting. My boot sequence automatically does a mysql repair on mythconverg.

I'm downloading R5F27 right now, on the theory that this has the schedules direct fix in it. When I tried to recompile mythtv just now with some additional logging of the startup of the backend added, the compile failed, saying I have a "hardware problem". Great.

So unless somebody has a magic bullet, I'll try upgrading to R5F27.

Bill

Only one thing I can think to try at this point - What happens if you restart the backend? /etc/init.d/mythtv-backend restart

We had someone else earlier this week who had mis-configured his BE/FE and networking so that they couldn't talk to each other...

In answer to your replies (and thanks)

1. restart is the same as "stop" followed by "start". Makes no difference,
the port 6543 is never opened.

2. I didn't mis-configure anything, I just replaced a dead power supply.

Anyway, it's all moot now, I've upgraded to R5F27.

Box is flakey, evidence suggests bad memory, so it's off to the store tomorrow morning for more memory.

Bill

The network identity problem could also occur if you were using DHCP with an unstable IP address and the DB or config files disagree.

if people would actually read the post, they would make fewer pointless suggestions.

STATIC IP stated in fourth paragraph in original post.

Bill

What you have to be careful of is that even though you only know of for sure that the power supply died, it could have taken some other components with it. You even suggest yourself that the appears to be a problem with memory.

Could I suggest that you download the ultimate boot CD http://ubcd.sourceforge.net/ and run some of the tests on this disk to check most of your components. You can test your memory and CPU and stuff like that.

I did the netstat command on my backend and I can see that it is listening. Could not think of much else, expect check the logs and see what the backend logs say about when it starts.

_________________
Girkers