First let me say that I've been running various versions of myth on this hardware for 4.5 years without issue. I just installed R5F27 and everything is working great except that for the problem described in the subject line.

I see the following error in the frontend log:

2007-10-18 22:30:32.022 MythSocket(b15c8610:18): readStringList: Error, timeout.
2007-10-18 22:30:32.048 Connection to backend server lost
2007-10-18 22:30:32.049 Connecting to backend server: 127.0.0.1:6543 (try 1 of 5)
2007-10-18 22:30:32.052 Using protocol version 31
2007-10-18 22:30:32.064 MythSocket(b15e8640:18): writeStringList: Error, invalid string list.

A search reveals many users having this problem, but no real solution has been suggested. I am not thrilled with running a cron job everyday to restart the processes.

Has anyone found a way to correct this problem?

does it not work at all, does it work for while then stop, was R5F27 a fresh install or an upgrade..

does "netstat -n | grep 6543" show anything.. do you have any iptables rules setup..

R5F27 was a fresh install.

Everything works greats for about 24 hours - web access, samba, recordings, frontend viewing, the works. Then the frontend loses connection.

I didn't touch iptables.
netstat looks fine

root@myth:/home/marty# netstat -n | grep 6543
tcp 0 0 127.0.0.1:6543 127.0.0.1:51451 ESTABLISHED
tcp 0 0 127.0.0.1:6543 127.0.0.1:51450 ESTABLISHED
tcp 0 0 127.0.0.1:51450 127.0.0.1:6543 ESTABLISHED
tcp 0 0 127.0.0.1:51451 127.0.0.1:6543 ESTABLISHED
tcp 0 0 127.0.0.1:38885 127.0.0.1:6543 ESTABLISHED
tcp 0 0 127.0.0.1:38884 127.0.0.1:6543 ESTABLISHED
tcp 0 0 127.0.0.1:6543 127.0.0.1:38885 ESTABLISHED
tcp 0 0 127.0.0.1:6543 127.0.0.1:38884 ESTABLISHED

This sounds strange, i know, but you could just do another fresh install to see if the problem is still there.. I did 4 fresh installs of R5F27 and every install had a different issue. After the last install everything worked great.

I did a reinstall. I still get the same problem. Everything still works perfectly except for the frontend losing connection to the backend on a daily basis.

For now, I'm working around it by programming a button on the remote to execute a script that restarts the frontend.

Anyone else have any ideas?

Is your frontend and backend one machine or two?

_________________
LinHES 8.3, 1 BE, 3 FE

One machine.

I have been having this exact same issue for months. I reinstalled today and the same thing seems to be happening.



This only happens when I leave a machine on the watch recordings screen. As you can see from above, for some reason it does not try to connect to the backend again. The backend is solid as a rock- gets all of my recordings and has never crasshed.

Heh, I almost wonder if there is a way to just kick out of this screen if I forget to exit to the menus....

My workaround is alos mapping a remote button to a frontend restart script- but it annoys me!

I found this link, but I hate to delete my whole database and start over:
http://www.gossamer-threads.com/lists/mythtv/users/291134?search_string=invalid%20string%20list;#291134

-Loco

For me, I believe that I solved my issues with losing connection. I got my information from this link:
http://knoppmythwiki.org/index.php?page=PerformanceOptimizeHowTo

I checked my database and optimized it. Then I reduced its size by limiting program data to 9 days instead of 14 as it recommended, by (just quoting the wiki here, I suggest checking out that link):


Maybe this will work for someone else with the problem?

-Loco