I made an update to your thread on SD. I think I should open a trouble ticket on the Myth bug trac, but I'm not sure. It looks like the schema upgrade code in mythtv-setup was supposed to handle ensuring the setting existed; I'm not sure why it didn't work, but I don't have the pre-upgrade setting either. I can't tell if the upgrade is loopy or if the KM installer may have done something. In either case, my personal, highly opinionated, opinion is that the UPDATE code should be more defensive and not rely on the pre-existing setting, but I'm not the owner of the code...

I suspect this impacts other distros, but the lack of activity on the SD thread seems to say otherwise. OTOH, this problem has no real impact or symptoms for the users. maybe no one really notices? My system has clearly been hitting SD 8x/day since forever. (My oldest backup is only from September, but the setting was missing there.) Users only care when the frequency gets below once per day or maybe twice per week. Schedules direct doesn't seem to be hurting from the extra runs. There's nothing actually 'broken' it is just extremely inefficient.

Note that this is after I disabled it again to keep it from continuing to run every 3 hours:

The housekeeping log shows the repeated run 3 hours after the time suggested above, before I caught it.

The log files from update_database.sh have rolled off due to some inconclusive manual debugging attempts so they're no longer available.

It looks like your MythFillSuggestedRunTime wasn't updated. It has 5:23 Housekeeping reports 8:27, then LastRun is 11:12

I don't have your update_database.sh or your settings.sh, so can't help there. I presume it's a wrapper for myhtfilldatabase...? am I missing something or is this a 6.0 feature?

if so, mythfilldatabase will only update the MythFillSuggestedRunTime if MythFillGrabberSuggestsTime is 1. so if you're debugging why the suggested run time isn't getting updated be sure to reset MythFillGrabberSuggestsTime before your run...

well, half-time is over...

Both have been published here in the past. The settings script is just a convenience wrapper for querying and updating the settings table. For the purposes here you could replace it with a "select * from settings". The update_database.sh script is fairly simple wrapper around mythfilldatabase. It does some logging, and some other db adjustments, but it's return code is the result from mythfilldatabase. Again it should be immaterial to the issue, it's been rock solid for ~2 years now, and hasn't even been touched since last April.

I have to take back my success report: I ended up getting the "grabber error" message on the MythWeb status page. I've reverted back to the workaround of having mythfilldatabase run at a given time range rather than the suggested time.

There must be multiple causes.[code]/var/log/mythtv$ grep mythfilldatabase mythbackend.log | tail -6
2009-01-30 04:08:59.953 Running mythfilldatabase
2009-01-30 07:11:01.037 Running mythfilldatabase
2009-01-30 10:13:01.486 Running mythfilldatabase
2009-01-31 14:28:49.373 Running mythfilldatabase
2009-02-01 22:06:40.678 Running mythfilldatabase
2009-02-02 06:27:06.915 Running mythfilldatabase[code]
I made my update around 8:00 1-30. Now 3 days hardly counts as "rock solid" I'll keep an eye on it...