PicorePlayer LMS Media scan terminated unexpectedly and then rescans randomly

[mherger]

> Just under 100k

143MB for 100k. OP has about double that number, or more. Which means
the file _could_ be 300MB.

It's possible that this really is very close to fitting the condition I
came up with at some point to decide over running in memory or not. I
guess LMS tries to optimize in memory, but actually doesn't have enough
memory, because my estimation isn't good enough.

Pommes - if you're reading here: how computer literate are you? Would
you feel comfortable editing a file or two to test whether we can figure
out what's going on?

You could also enable INFO logging for database.info
(Settings/Advanced/Logging) before re-running a scan with max memory
mode. That would give us some information about what's going on without
modifying the code. Do NOT set the log level to DEBUG, as this could
create a huge scanner.log!

--

Michael

[mherger]

> 357.6M library.db

Yeah, this could really be just a bit too much for in-memory
optimization, but not large enough to trigger my condition. Please see
my previous response to Paul.

--

Michael

[Pommes]

> Just under 100k

143MB for 100k. OP has about double that number, or more. Which means
the file _could_ be 300MB.

It's possible that this really is very close to fitting the condition I
came up with at some point to decide over running in memory or not. I
guess LMS tries to optimize in memory, but actually doesn't have enough
memory, because my estimation isn't good enough.

Pommes - if you're reading here: how computer literate are you? Would
you feel comfortable editing a file or two to test whether we can figure
out what's going on?

You could also enable INFO logging for database.info
(Settings/Advanced/Logging) before re-running a scan with max memory
mode. That would give us some information about what's going on without
modifying the code. Do NOT set the log level to DEBUG, as this could
create a huge scanner.log!

--

Michael

Well, i am not that IT professionell, all the commands i use i found out on google:-)
But if you´d give me step by step instructions, i would be pleased to help.
I am very grateful for this community here, so i will do all i can:-)

[paul-]

pCP 32 bit will use all memory of the rpi4 boards, but its 3GB per process. (2GB in practicality with overhead) There is however a full 64bit pCP that doesn't have these restrictions.

pCP7 is in beta now, I'm waiting for a few kernel fixes from RPI (for some fridge case uses - not normal LMS/squeezelite), then I'll push out the normal release.

[bpa]

pCP 32 bit will use all memory of the rpi4 boards,

even the 8Gb boards ?

[paul-]

Yes. They are using a LPAE in the kernels. So there is some overhead to do it.

[mherger]

>> You could also enable INFO logging for database.info
>> (Settings/Advanced/Logging) before re-running a scan with max memory
>> mode. That would give us some information about what's going on without
>> modifying the code. Do NOT set the log level to DEBUG, as this could
>> create a huge scanner.log!
>
> Well, i am not that IT professionell, all the commands i use i found out
> on google:-)
> But if you´d give me step by step instructions, i would be pleased to
> help.

Please try the above with memory set to max. Yes I know, it'll crash the
scanner. But that's what I'm looking for. Maybe if you could monitor the
memory usage when it's getting closer to the optimize stage (and I know
again, this asking much :-)).

--

Michael

[Pommes]

>> You could also enable INFO logging for database.info
>> (Settings/Advanced/Logging) before re-running a scan with max memory
>> mode. That would give us some information about what's going on without
>> modifying the code. Do NOT set the log level to DEBUG, as this could
>> create a huge scanner.log!
>
> Well, i am not that IT professionell, all the commands i use i found out
> on google:-)
> But if you´d give me step by step instructions, i would be pleased to
> help.

Please try the above with memory set to max. Yes I know, it'll crash the
scanner. But that's what I'm looking for. Maybe if you could monitor the
memory usage when it's getting closer to the optimize stage (and I know
again, this asking much :-)).

--

Michael

Yes i can do that, a complete delete and rescan everything?
How would i monitor the memory usage?

[paul-]

open up a shell and run the command `top`



watch the MiB Mem: line.

[mherger]

> Yes i can do that, a complete delete and rescan everything?

SSH in to the device
(https://docs.picoreplayer.org/how-to...s_pcp_via_ssh/), run "top".
The top line would give you something like this:

Mem: 222824K used, 775996K free, 13088K shrd, 17652K buff, 68236K cached

Just try to see whether during the optimize step the usage increases a
lot. I would expect it to.

--

Michael