or .
connect with an HDMI cable to the TV that the RPi sits next to, do
Code:
sudo shutdown -r now
via SSH for a reboot and then just change network via Raspbian
interface on the TV.
Now I remember how I did it in the first place - doh!
Actually
I guess I got it working. I built a chroot environment on my laptop for
qemu emulation thinking the build would go quicker, wow, it was
sloowww, much quicker compiling directly on the pi.
I don't know if everthing is really working correctly quite yet but
squeezeboxserver starts:
Code:
--
So either the update is failing to install or the pop-up notification is
faulty. Advice? Thanks.
There's an updates sub-folder in the LMS cache (see
Settings/Information). Delete the installer and/or server.version files
from in there.
--
Michael
The host is a Supermicro A1SAi-2750F with an Atom C2750 + 16GB Ram
Currently, the library consists of 188.570 titles.
Ok, that's not a small collection. But the CPU isn't poor either.
When using the search box, every 1 in 10 search queries only returns
results after ~20-30s
What exact search
The host is a Supermicro A1SAi-2750F with an Atom C2750 + 16GB Ram
Currently, the library consists of 188.570 titles.
Ok, that's not a small collection. But the CPU isn't poor either.
When using the search box, every 1 in 10 search queries only returns
results after ~20-30s
What exact search
Yatsushiro wrote:
> Is it important that your brothers server runs on Fedora?
>
> If not, have you considered running PcP as an LMS server, or Ubuntu
> server? I run the latter on a Pi4, no issues, great performance.
Hi Yatsushiro, no, not especially important but I run it on my server
and lap
bpa wrote:
>
> Have you tried the perfwarn command line setting in case it might
> confirm where the code hot spot is ?
Roland0 wrote:
> Switch logging to info in Settings>Advanced>Logging for all database.*
> entries and check the log after a slow query
I'm trying both now. We'll see in a
Hypfer wrote:
> Nope. No patterns nor different complexity in the search terms. One that
> causes this issue will work perfectly fine three days later while one
> that works fine today might show these issues tomorrow.
>
Switch logging to info in Settings>Advanced>Logging for all database.*
en
On the Mac (macOS 10.14.6, Mojave) the LMS System Preferences tells me
daily there's an updated version available, even if there apparently
isn't. I'm already at LMS 7.9.2r1575296326.
The installer reports that it was successful. And I'm left at the same
rev level.
So either the update is fail
Hypfer wrote:
> It's also 100% not IO related. I've checked that multiple times on
> multiple occasions. It's always just one CPU core at 100% actually busy
> doing something. Not iowait
The point I was making is that if an read/write request (e.g. there are
a few cases in LMS code/plugins where
d6jg wrote:
> Nice simple system.
> De-bugging that ought to be really easy (not).
>
> You said the issue occurs perhaps 1 in 10 searches - is there anything
> different / more complex about the 1 i.e. might return more results than
> the other 9
> Why a docker container? is it possible to run
What is the version of the IO::Socket:SSL ?
Just to take a different view.
A consistent 20-30 sec delay "feels" like a network timeout.
Could LMS be "spinning" on the scheduler select ?
A while back there was a case of https socket blocking - wonder if there
is something similar.
-
Hypfer wrote:
> Yes
>
>
>
> The host is a Supermicro A1SAi-2750F with an Atom C2750 + 16GB Ram
> LMS 7.9.2 - 1574315728 is running in a docker container. The database is
> located in a ZFS dataset on a ZFS Mirror Pool utilizing two sufficient
> Sata SSDs.
> Music is stored in another ZFS data
bpa wrote:
> You have enabled the Database Memory Config as "High" for > 1Gbyte RAM
Yes
mherger wrote:
> > Thats not an option here because the host is quite capable due to its
> > eight cores.
> > It's just the single-core performance which is below average
>
> What exactly are we talking a
sodface wrote:
>
> The build keeps failing on an include of xlocale.h in digitlst.cpp and
> from what I can gather, xlocale.h has been removed from Fedora for a
> while. I manually changed it to just locale.h and restarted build.sh
> then it failed again with the same error. Hmmm, I thought m
However, OP doesn't specify which DB queries take that long, so
it might be a plugin etc.
Exactly why I believe we need more information before we can suggest
anything useful.
This sadly causes quite noticable performance issues (15s+ delays on DB
Queries)
so that's what my answer was adre
Is it important that your brothers server runs on Fedora?
If not, have you considered running PcP as an LMS server, or Ubuntu
server? I run the latter on a Pi4, no issues, great performance.
Yatsushiro's Profile: http://f
17 matches
Mail list logo