Opened 22 months ago

Last modified 11 months ago

#12994 new Bug Report - General

On 32bit systems with qt 5.8 mythbackend permanently on 100%CPU load, 75% system memory

Reported by: jorohr@… Owned by:
Priority: minor Milestone: 29.2
Component: MythTV - General Version: Master Head
Severity: medium Keywords:
Cc: Ticket locked: no

Description

[Apologies for duplicates, I could find the bug with the same title I submitted earlier today]

I am running mythtv on a Raspberry Pi 3 on Arch Linux Arm (armv7h). Since the last upgrade, the mythbackend process permanently caused 100% cpu load and used 75% system memory. When a client tried to connect, it was mostly unresponsive.

Today I downgraded the QT packages from 5.8 to 5.7 and things were back to normal. The same has been reported for the Banana Pi, see https://archlinuxarm.org/forum/viewtopic.php?f=15&t=11278&sid=59d9bfe7685e9f81ee5eb4eacb490617

Change History (12)

comment:1 Changed 22 months ago by Peter Bennett

The highest version of QT available on Ubuntu is QT 5.7 on the pre-release Ubuntu 17.04 version, so it is likely that MythTV has not been tested with 5.8 on any platform. It is possible that this problem not exclusive to raspberry pi or banana pi.

comment:2 Changed 22 months ago by Stuart Auchterlonie

Milestone: unknown29.0
Version: UnspecifiedMaster Head

comment:3 Changed 22 months ago by brendan@…

For what it's worth, compiling Myth 0.28.1 fails with QT5.8. Compiling with QT5.7 is fine though.

comment:4 in reply to:  3 Changed 21 months ago by Stuart Auchterlonie

Replying to brendan@…:

For what it's worth, compiling Myth 0.28.1 fails with QT5.8. Compiling with QT5.7 is fine though.

That's been fixed.

Regards Stuart

comment:5 Changed 21 months ago by brendan@…

I was having the same failure on an Intel Atom system with Arch Linux i686 (32bit). Running with QT5.7 was fine but CPU usage went up to 100% on QT5.8 before the backend failed with a segfault. Today, I upgraded the same system to x86_64 with QT5.8 and the backend runs fine. This would seem to imply that the issue is with the 32 bit version of QT5.8.

Regards

Brendan

comment:6 in reply to:  5 ; Changed 20 months ago by wilberfan <wilberfan@…>

Replying to brendan@…:

I was having the same failure on an Intel Atom system with Arch Linux i686 (32bit). Running with QT5.7 was fine but CPU usage went up to 100% on QT5.8 before the backend failed with a segfault. Today, I upgraded the same system to x86_64 with QT5.8 and the backend runs fine. This would seem to imply that the issue is with the 32 bit version of QT5.8.

Similar setup here (Arch Linux, 32-bit), and I'm still having this issue even after updating to the latest mythtv and QT5.8. (Dual-core Intel has 80% CPU on both cores from mythbackend with entire system lagging.) Reverting to QT5.7 restores functionality. This would seem to confirm there's still an issue with 32-bit QT5.8?

comment:7 Changed 20 months ago by Peter Bennett

Summary: On armv7h (raspberrypi3, banana pi) with qt 5.8 mythbackend permanently on 100%CPU load, 75% system memoryOn 32bit systems with qt 5.8 mythbackend permanently on 100%CPU load, 75% system memory

I changed the title because it does not seem to be related to armv7h (raspberrypi3, banana pi)

comment:8 in reply to:  6 Changed 20 months ago by wilberfan <wilberfan@…>

Replying to wilberfan <wilberfan@…>:

Replying to brendan@…:

I was having the same failure on an Intel Atom system with Arch Linux i686 (32bit). Running with QT5.7 was fine but CPU usage went up to 100% on QT5.8 before the backend failed with a segfault. Today, I upgraded the same system to x86_64 with QT5.8 and the backend runs fine. This would seem to imply that the issue is with the 32 bit version of QT5.8.

Similar setup here (Arch Linux, 32-bit), and I'm still having this issue even after updating to the latest mythtv and QT5.8. (Dual-core Intel has 80% CPU on both cores from mythbackend with entire system lagging.) Reverting to QT5.7 restores functionality. This would seem to confirm there's still an issue with 32-bit QT5.8?

Just received a notice that QT5.9-beta has been released. Not sure how that would affect debugging this issue. https://blog.qt.io/blog/2017/04/07/qt-5-9-beta-released/

comment:9 Changed 17 months ago by brendan@…

I recently tried again and compiled with QT5.9. The CPU load problem went away so I assume that it was just a QT5.8 issue?

comment:10 Changed 13 months ago by Stuart Auchterlonie

Milestone: 29.029.1

comment:11 Changed 11 months ago by Stuart Auchterlonie

Milestone: 29.10.28.2

Moving remaining open tickets to 0.28.2 milestone

comment:12 Changed 11 months ago by Stuart Auchterlonie

Milestone: 0.28.229.2

Moving remaining open tickets to 29.2 milestone

Note: See TracTickets for help on using tickets.