Project

General

Profile

Crash without any trace

Added by Zenon Skuza over 7 years ago

I'm running tvheadend 4.1-2484~g1fddf329a via a docker (linuxserver/tvheadend) on unRAID and getting crashes after recording about 3+ hours of 720p TV.

I suspect the docker is hitting RAM limits as I can see the docker's memory allocation just continually rising the moment recordings start, by 2MB/sec - long recordings may be hitting a limit.

I set the following debug traces on, but at the point of tvheadend crash, there is no log message of note

main,CPU,parser,TS,muxer,pass,mkv,subscription,profile,dvr,dvb,mpegts,linuxdvb

Are there any other debug traces that might assist with memory allocation/usage or similar potential crash causes?


Replies (3)

RE: Crash without any trace - Added by Mark Clarkstone over 7 years ago

Zenon Skuza wrote:

I'm running tvheadend 4.1-2484~g1fddf329a via a docker (linuxserver/tvheadend) on unRAID and getting crashes after recording about 3+ hours of 720p TV.

I suspect the docker is hitting RAM limits as I can see the docker's memory allocation just continually rising the moment recordings start, by 2MB/sec - long recordings may be hitting a limit.

I set the following debug traces on, but at the point of tvheadend crash, there is no log message of note

main,CPU,parser,TS,muxer,pass,mkv,subscription,profile,dvr,dvb,mpegts,linuxdvb

Are there any other debug traces that might assist with memory allocation/usage or similar potential crash causes?

Change the recording profile to sync + don't keep.

RE: Crash without any trace - Added by Zenon Skuza over 7 years ago

Thanks Mark,

That seems to have done the trick. Sitting around 44Mb mem usage regardless of number of recording tasks. Will keep monitoring to see how it goes.

Is this a bug in the System recording profile or just a problem with the tvheadend docker I'm using?

RE: Crash without any trace - Added by Mark Clarkstone over 7 years ago

Zenon Skuza wrote:

Thanks Mark,

That seems to have done the trick. Sitting around 44Mb mem usage regardless of number of recording tasks. Will keep monitoring to see how it goes.

Is this a bug in the System recording profile or just a problem with the tvheadend docker I'm using?

I honestly couldn't tell you, I've used system on my own machine (non-docker instance) & it's been fine. The same problem occurs on the RPi. May just be tvheadend being greedy with RAM, the system saying no & tvheadend throwing a wobblier as a result.

    (1-3/3)