Recording not started unless mythtv-backend is restarted

[Klavs Klavsen <kl@…>]

I'm using mythtv-0.22+fixes (svn 20091118) - on a 1200mhz, 512mb ram Armel arch. I build it as a package, based on Ubuntu karmic's 0.22+fixes rules minus the patches.

I'll gladly share them, if you want packages for debian Lenny armel arch.

I have set mythtv up to record several shows, but I just noticed that a show it was supposed to record at 20:00 - it hadn't started to record (it has otherwise started recordings as needed) and mythweb and mythfront end couldn't get any answers from the backend.

I restarted the backend at 20:24 - and it immediately started recording (24 minutes too late :( )

Here's the log: 2009-11-23 19:03:25.439 Preview: Grabbed preview '/home/mythtv/storage/1002_20091123170500.mpg' 704x576@360s 2009-11-23 19:32:31.344 UPnpMedia: BuildMediaMap? - no VideoStartupDir? set, skipping scan. 2009-11-23 20:02:36.432 UPnpMedia: BuildMediaMap? - no VideoStartupDir? set, skipping scan. 2009-11-23 20:24:41.684 mythbackend version: branches/release-0-22-fixes [svn20091118] www.mythtv.org 2009-11-23 20:24:41.695 Using runtime prefix = /usr 2009-11-23 20:24:41.700 Using configuration directory = /var/lib/mythtv/.mythtv 2009-11-23 20:24:41.758 Empty LocalHostName?. 2009-11-23 20:24:41.760 Using localhost value of debian 2009-11-23 20:24:42.005 New DB connection, total: 1 2009-11-23 20:24:42.045 Connected to database 'mythconverg' at host: localhost 2009-11-23 20:24:42.047 Closing DB connection named 'DBManager0' 2009-11-23 20:24:42.052 Connected to database 'mythconverg' at host: localhost 2009-11-23 20:24:42.106 Current MythTV Schema Version (DBSchemaVer): 1244 2009-11-23 20:24:42.118 MythBackend?: Starting up as the master server. 2009-11-23 20:24:42.180 New DB connection, total: 2 2009-11-23 20:24:42.183 Connected to database 'mythconverg' at host: localhost 2009-11-23 20:24:42.282 New DB connection, total: 3 2009-11-23 20:24:42.285 Connected to database 'mythconverg' at host: localhost

2009-11-23 20:24:44.083 New DB scheduler connection 2009-11-23 20:24:44.086 Connected to database 'mythconverg' at host: localhost 2009-11-23 20:24:44.291 Enabling Upnpmedia rebuild thread. 2009-11-23 20:24:45.701 Main::Registering HttpStatus? Extension 2009-11-23 20:24:45.702 Enabled verbose msgs: important general 2009-11-23 20:24:45.739 AutoExpire?: CalcParams?(): Max required Free Space: 1.0 GB w/freq: 15 min 2009-11-23 20:24:47.166 Reschedule requested for id -1. 2009-11-23 20:24:50.209 Scheduled 86 items in 3.0 = 0.11 match + 2.87 place 2009-11-23 20:24:50.225 AUTO-Startup assumed 2009-11-23 20:24:50.262 TVRec(1): ASK_RECORDING 1 0 0 0 2009-11-23 20:24:50.466 TVRec(1): Changing from None to Watching RecordingOnly? 2009-11-23 20:24:50.475 TVRec(1): HW Tuner: 1->1 2009-11-23 20:24:50.517 TVRec(3): ASK_RECORDING 3 0 0 0 2009-11-23 20:24:50.583 AutoExpire?: CalcParams?(): Max required Free Space: 3.0 GB w/freq: 15 min 2009-11-23 20:24:50.589 Started recording: Hvem vil være millionær?: channel 1002 on cardid 1, sourceid 1 2009-11-23 20:24:54.298 UPnpMedia: BuildMediaMap? - no VideoStartupDir? set, skipping scan.

I checked (before restarting) if the box was overloaded somehow, since I didn't get an answer, but the box was idle: # vmstat 1 procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----

r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 48304 5168 1348 457048 0 0 21 20 18 3 12 2 85 0 0 0 48304 5168 1348 457048 0 0 0 0 216 184 2 0 98 0 0 0 48304 5168 1348 457048 0 0 0 0 218 192 2 1 97 0

Note that there is an entry right after 20:00.. - but it didn't start recording even though a small part of myth-backend was alive.

The backend has run for some time - so I don't think I can run it with gdb running all the time - but perhaps gdb can be attached to a running process - or perhaps I can enable debug log output to help solve this issue?

[Klavs Klavsen <kl@…>]

It happened again this morning/night.

It even "asked" for recording -but nothing happened (that's how I interpret the log): 2009-11-24 06:55:52.167 UPnpMedia: BuildMediaMap? - no VideoStartupDir? set, skipping scan. 2009-11-24 06:56:16.066 AutoExpire?: CalcParams?(): Max required Free Space: 1.0 GB w/freq: 15 min 2009-11-24 07:01:11.022 Reschedule requested for id -1. 2009-11-24 07:01:14.853 Scheduled 94 items in 3.8 = 0.56 match + 3.20 place 2009-11-24 07:04:52.471 Reschedule requested for id -1. 2009-11-24 07:04:56.296 Scheduled 94 items in 3.8 = 0.55 match + 3.21 place 2009-11-24 07:07:39.960 Reschedule requested for id -1. 2009-11-24 07:07:47.697 Scheduled 94 items in 7.3 = 0.74 match + 6.55 place 2009-11-24 07:08:00.721 Reschedule requested for id 0. 2009-11-24 07:08:04.573 Scheduled 94 items in 3.8 = 0.00 match + 3.78 place 2009-11-24 07:08:29.639 TVRec(2): ASK_RECORDING 2 29 0 0 2009-11-24 07:08:29.858 TVRec(7): ASK_RECORDING 7 29 0 0 2009-11-24 07:25:53.656 UPnpMedia: BuildMediaMap? - no VideoStartupDir? set, skipping scan. 2009-11-24 07:55:55.154 UPnpMedia: BuildMediaMap? - no VideoStartupDir? set, skipping scan. 2009-11-24 08:26:02.162 UPnpMedia: BuildMediaMap? - no VideoStartupDir? set, skipping scan. 2009-11-24 08:56:08.169 UPnpMedia: BuildMediaMap? - no VideoStartupDir? set, skipping scan. 2009-11-24 08:57:25.747 mythbackend version: branches/release-0-22-fixes [svn20091118] www.mythtv.org 2009-11-24 08:57:25.788 Using runtime prefix = /usr 2009-11-24 08:57:25.790 Using configuration directory = /var/lib/mythtv/.mythtv

[Klavs Klavsen <kl@…>]

Would it help if I attached to the process with gdb - when I discover it has died?

Using this guide: ​https://wiki.ubuntu.com/Backtrace#Already%20running%20programs

[anonymous]

Hello.

I'm having the same issue.

I have been running MythTV under Gentoo for over 5 years over many versions, however this is the first time I've been running a completely separate backend. I now have three minimyth front ends on a variety of hardware from an AMD64 to a Pentium 4 to a newer ATOM.

I am also running heyu, which is an X10 utility, plus squeezeboxserver for music, as well as apache2. cpudyn is taking care of CPU power and hard disk spindown on my main disk.

Everything works great for about 2 to 3 week after a fresh backend boot, and then suddenly, one day, mythbackend stops recording. Everything else is working, including the ability to see upcoming recordings, but the backend will not simply record.

Beside the conversion to running only mythbackend on a dedicated server, I upgraded to mythtv ver 22824 when this occurred and added a HDHomeRun capture device during the same upgrade. minimyth was also upgraded to version 0.22.0-72 which was the latest stable release.

The server is an SiS 764 chipset board with a Turion64 processor. There are 2 PVR500 boards in it, for 4 analog capture. There is 2GB of memory and 4 GB of swap available. The only thing that seems to be unusual is I have the commercial flagging set to occur while recording is in progress, rather than once the entire program has finished recording. This means a top will show mythcommflag taking up to 80 - 90 % of the CPU. However mythcommflag runs as priority 37, which is the lowest priority task.

[danielk]

Marking as invalid mostly because of the age of this ticket, if it were fresher I'd ask for a backtrace of the backend when it appears wedged. If this happens with 0.23 please create a backtrace by attaching to the backend process and then typing "thread apply all bt" and of course the log from when the backend is wedged would be useful.