I recently decided to pull the trigger and do a completely rebuild of my old myth-backend (only overlapping hardware is case, PS, and a 4TB movie storage drive). OS is Ubunutu 20.04 LTS. [I miss the ease of mythbuntu....]
While tracking thermal performance (I plan on downsizing the case), I noticed an odd behavior. Every ~1min, 1 core of my CPU (4-core) will jump up to100% for 10-15sec. This heats up the 1 core up, ~30C to 50C+. The culprit thread is /usr/sbin/mysqld. The behavior stops when I stop mythbackend.
While looking at the mythtbackend log, I see this repeating every 1min:
<log file>
Code: Select all
May 5 06:53:38 rickter-htpc mythbackend: mythbackend[17332]: I CoreContext housekeeper.cpp:754 (Run) Queueing HouseKeeperTask 'MythFillDB'.
May 5 06:53:38 rickter-htpc mythbackend: mythbackend[17332]: I HouseKeeping housekeeper.cpp:144 (Run) Running HouseKeeperTask 'MythFillDB'.
May 5 06:53:39 rickter-htpc mythbackend: mythbackend[17332]: I ProcessRequest mainserver.cpp:1768 (HandleAnnounce) MainServer: MainServer::ANN Playback
May 5 06:53:39 rickter-htpc mythbackend: mythbackend[17332]: I ProcessRequest mainserver.cpp:1770 (HandleAnnounce) MainServer: adding: rickter-htpc(55ab4c447cc0) as a client (events: 0)
May 5 06:53:58 rickter-htpc mythbackend: mythbackend[17332]: I Scheduler scheduler.cpp:2308 (HandleReschedule) Reschedule requested for MATCH 0 0 0 - MythFillDatabase
May 5 06:53:58 rickter-htpc mythbackend: mythbackend[17332]: I Scheduler scheduler.cpp:2425 (HandleReschedule) Scheduled 67 items in 0.0 = 0.02 match + 0.02 check + 0.01 place
May 5 06:53:58 rickter-htpc mythbackend: mythbackend[17332]: I MythSocketThread(44) mainserver.cpp:7856 (connectionClosed) Playback sock(55ab4c447cc0) 'rickter-htpc' disconnected
May 5 06:53:58 rickter-htpc mythbackend: mythbackend[17332]: I HouseKeeping housekeeper.cpp:164 (Run) HouseKeeperTask 'MythFillDB' Finished Successfully.
Is this a bug in the MythFillDB task?? Does anyone else using Myth31 see this behavior? To be honest, if wasn't looking at thermal performance so closely, I likely would've missed it.
[btw -- nothing out of the ordinary in syslog]
Thanks,
-Rick