mamema wrote:
> have this in the logs today:
>
> [21-04-29 09:51:04.8337] Slim::Networking::Repositories::__ANON__ (147)
> Failed to fetch https://www.nexus0.net/pub/sw/lmsdboptim/repo.xml:
> Connect timed out: Inappropriate ioctl for device
> [21-04-29 09:51:04.8340]
have this in the logs today:
[21-04-29 09:51:04.8337] Slim::Networking::Repositories::__ANON__ (147)
Failed to fetch https://www.nexus0.net/pub/sw/lmsdboptim/repo.xml:
Connect timed out: Inappropriate ioctl for device
[21-04-29 09:51:04.8340] Slim::Plugin::Extensions::Plugin::_noResponse
(476)
SpiderJon wrote:
> I ran
>
> find /usr/share/squeezeboxserver -name SQLite.so
>
> which gave
>
>
/usr/share/squeezeboxserver/CPAN/arch/5.20/arm-linux-gnueabi-thread-multi-64int/auto/DBD/SQLite/SQLite.so
>
Roland0 wrote:
> Which perl version are you using?
>
> Find your current module:
> >
Code:
> > find /usr/share/squeezeboxserver -name SQLite.so
>
> >
> and then run (substituting your path)
> >
Code:
> > file
I've just published a new version. There are quite a number of changes,
so if you notice something, feel free to provide feedback.
'Various SW' (https://www.nexus0.net/pub/sw/): Web Interface | TUI |
Playlist Editor / Generator | Music Classification | Similar Music |
Announce | EventTrigger
mherger wrote:
> >
> Ah, no, not with the docker image. You'd have to build your own (based
> on ours), or run LMS without docker or whatever.
No hat was not the intention of this test, so we're staying with sqlite
then.
there is no MySQL driver around.installing it? Hmm, its docker...
you are maintaining the docker image, sure for fun to test, but you
still sure it should work?
Ah, no, not with the docker image. You'd have to build your own (based
on ours), or run LMS without docker or whatever.
mherger wrote:
> > this doesn't seem to be the case with 8.2.
> [/color]
>
> Double/triple check the connection string. Setting LMS up to use MySQL
> might be fiddly. But it should definitely work to a certain degree.
> Meaning there might be bugs around the SQL dialect being used or
>
this doesn't seem to be the case with 8.2.
It certainly still should be possible.
I've created the database in mysql, created the suitable user (according
to server.prefs), granted rights to the database, changed dbsource entry
to mysql according to perl notation, i even flushed the cache
mherger wrote:
> >
> But you should be able to use MySQL, if you wanted to give it a try.
> Would be interesting if that would turn out to be faster. One advantage
>
> certainly would be that it would use another core on today's multi-core
>
> systems, whereas SQLite is running inside the
SpiderJon wrote:
> Yes, please. I'm running Debian 10 (Buster) on a Pi4 (armv7l-linux) so
> 32bit Arm.
Which perl version are you using?
Find your current module:
Code:
find /usr/share/squeezeboxserver -name SQLite.so
and then run
mherger wrote:
> >> Mine is 440Mb - with around 25k artists although not 100% good tags
> and
> >> many obscure artists that probably have no image.
> > While my library only has ~5000 artists, this still doesn't account
> for
> > the DB size difference.
>
> Haha... I only have a few hundred
Roland0 wrote:
> It's part of the perl module DBD::SQLite's source code, which
> unfortunately means you have to re-compile it.
> On the plus side, it only takes a few minutes and you get a shiny new
> SQLite version for free.
>
> Or, if your LMS is running on 32bit Arm or 64bit x86, I could
Mine is 440Mb - with around 25k artists although not 100% good tags and
many obscure artists that probably have no image.
While my library only has ~5000 artists, this still doesn't account for
the DB size difference.
Haha... I only have a few hundred album artists!
What are your exact MAI
i will, as i said, it's pure fun for me.
As you seems to be very knowlegeable with databases, do you have
suggestions to test the performance in a more "scientific" way, than
"seems faster"..?
You could write a script to run a bunch of requests: get a list of x
albums, get the artwork for
mamema wrote:
> i will, as i said, it's pure fun for me.
> As you seems to be very knowlegeable with databases, do you have
> suggestions to test the performance in a more "scientific" way, than
> "seems faster"..?
>
As luck would have it, 'there actually is a benchmark'
Roland0 wrote:
> While my library only has ~5000 artists, this still doesn't account for
> the DB size difference.
> Do you have " Find artwork online during library scan" enabled (I
> don't)?
>
> Whatever the reason, it seems to make sense to add imageproxy.db to the
> DB supported by the
Roland0 wrote:
>
> Nevertheless, I agree, if anyone bothers to invest time in actually
> benchmarking this, it'll interesting to the see the results.
>
i will, as i said, it's pure fun for me.
As you seems to be very knowlegeable with databases, do you have
suggestions to test the
mherger wrote:
> > I am (artists only), but my imageproxy.db is only 8MB (even though
> I've
> > checked " Download and pre-cache artist pictures from online
> sources"),
> > so I haven't bothered with it.
> > Do you have an idea what typical sizes would be?
>
> Hmm... mine is about 770MB!
Paul Webster wrote:
> Mine is 440Mb - with around 25k artists although not 100% good tags and
> many obscure artists that probably have no image.
While my library only has ~5000 artists, this still doesn't account for
the DB size difference.
Do you have " Find artwork online during library
I am (artists only), but my imageproxy.db is only 8MB (even though I've
checked " Download and pre-cache artist pictures from online sources"),
so I haven't bothered with it.
Do you have an idea what typical sizes would be?
Hmm... mine is about 770MB! Pre-caching actually should be using it.
mamema wrote:
> i'm on of those guys willing to throw more RAM on the issue. I'm also
> willing to test (with you) how much more RAM is really needed to speed
> up the things. I have 32 GB which are more or less usable. But i also
> think, there must be a break even point where adding more RAM
Shozzer wrote:
> II did a complete clear and rescan but funnily enough that seemed to
> take longer. I will have to keep an eye on that.
>
I'm moderately confident that the plugin's DB settings do not adversely
affect performance (unless the system runs out of memory, but in this
case, the DB
SpiderJon wrote:
>
> Re: Re-enabling the sqlite_stat4 table in "Other methods to improve
> performance" on your 'LMS DB Optimizer info page'
> (https://www.nexus0.net/pub/sw/lmsdboptim/).
> Where is Makefile.PL? And is there anything else you need to do after
> uncommenting the line
Roland0 wrote:
> I am (artists only), but my imageproxy.db is only 8MB (even though I've
> checked " Download and pre-cache artist pictures from online sources"),
> so I haven't bothered with it.
> Do you have an idea what typical sizes would be?
Mine is 440Mb - with around 25k artists
mherger wrote:
>
> But you should be able to use MySQL, if you wanted to give it a try.
> Would be interesting if that would turn out to be faster. One advantage
>
> certainly would be that it would use another core on today's multi-core
>
> systems,
>
I'd speculate MySQL won't be faster.
mherger wrote:
> > I'm still fiddling with some of the parameters (cache sizes, mainly).
> > Currently, they should be fine for ~1-2GB free memory.
>
> I'll have to play with this, in particular with the imageproxy.db, too:
>
> MAI makes heavy use of it to cache artist artwork etc... Are you
mherger wrote:
> >
> But you should be able to use MySQL, if you wanted to give it a try.
> Would be interesting if that would turn out to be faster. One advantage
>
> certainly would be that it would use another core on today's multi-core
>
> systems, whereas SQLite is running inside the
Shozzer wrote:
> I'm not sure that sugar enters into it. More like steroids. When
> TrackStat was enabled for LMS 8.* I immediately installed it, only to
> uninstall it very shortly afterwards - starting and stopping LMS was
> taking too long. Out of curiosity I have installed the Optimizer
philchillbill wrote:
> How much less sugar does it have than the regular non-diet version ?
> [emoji23]
>
>
> Sent from my iPhone using Tapatalk
I'm not sure that sugar enters into it. More like steroids. When
TrackStat was enabled for LMS 8.* I immediately installed it, only to
uninstall it
Roland0 wrote:
> DB Optimizer is a plugin for LMS to improve it's database performance
> and memory use.
> More information can be found 'on the homepage'
> (https://www.nexus0.net/pub/sw/lmsdboptim/).
Hi Roland0
Re: Re-enabling the sqlite_stat4 table in "Other methods to improve
Also, because i'm curious and have DB spcialists here, why is the whole
stuff file based and would a dedicated DB server or a Redis server help?
File based: it's simpler and good enough.
We used to ship with MySQL for a while, but maintaining MySQL in the
background on Windows in a user
Roland0 wrote:
> Thanks for the feedback.
>
> I'm still fiddling with some of the parameters (cache sizes, mainly).
> Currently, they should be fine for ~1-2GB free memory.
> However, if enough people with different requirements use this (probably
> unlikely...), it might make sense to add
Shozzer wrote:
> Fair enough, that makes sense. I must be running SQLite!
How much less sugar does it have than the regular non-diet version ?
[emoji23]
Sent from my iPhone using Tapatalk
philchillbill's Profile:
mherger wrote:
> > I am running LMS 8.1.1 on Windows 10. How do I ascertain if I am
> using
> > MySQL or SQLite? There are folders for both within Squeezebox\server.
>
> If you don't know what you're using, you're using SQLite :-)
>
> MySQL really only is there for legacy reason. There might
I am running LMS 8.1.1 on Windows 10. How do I ascertain if I am using
MySQL or SQLite? There are folders for both within Squeezebox\server.
If you don't know what you're using, you're using SQLite :-)
MySQL really only is there for legacy reason. There might be some users
of it, but it
Hi
> Only SQLite is supported, don't install this if your LMS uses MySQL.
I am running LMS 8.1.1 on Windows 10. How do I ascertain if I am using
MySQL or SQLite? There are folders for both within Squeezebox\server.
Thanks.
Steve
I'm still fiddling with some of the parameters (cache sizes, mainly).
Currently, they should be fine for ~1-2GB free memory.
I'll have to play with this, in particular with the imageproxy.db, too:
MAI makes heavy use of it to cache artist artwork etc... Are you using MAI?
philchillbill wrote:
> I just updated LMS on my Ubuntu NUC to 8.2 and also added your plugin.
> The UI feels much snappier, especially in iPeng with loading thumbnails
> for cover art in the New Music view. Nice work !
>
> A new and changed rescan of a DB with 70K tracks took 42 seconds. The
>
I just updated LMS on my Ubuntu NUC to 8.2 and also added your plugin.
The UI feels much snappier, especially in iPeng with loading thumbnails
for cover art in the New Music view. Nice work !
A new and changed rescan of a DB with 70K tracks took 42 seconds. The
system has an SSD but the music
Roland0 wrote:
> 'gotop' (https://github.com/xxxserxxx/gotop/releases) or 'bottom'
> (https://github.com/ClementTsang/bottom/releases) have binaries for most
> platforms. You could add one of them to your container.
neither of both are working on Synology with the provided binaries. So
i've
Roland0 wrote:
> Could very well be - it's certainly incorrect, even though my local
> installation didn't complain.
> Anyway, I've uploaded a new .zip
That seems to have worked. I guess it was a cache problem.
Thank you
2 x Touch
2 x Radio
2 x Boom
1 x Intel-NUC server/squeezelite running
last version working for me at least, on debian buster. many thanks for
this :D
manx's Profile: http://forums.slimdevices.com/member.php?userid=70875
View this thread: http://forums.slimdevices.com/showthread.php?t=114067
bobertuk wrote:
> Still the same. I noticed that
> C:\ProgramData\Squeezebox\Cache\InstalledPlugins\Plugins\LMSdbopt/install.xml
> contains 0.0.1
> Could that be the problem?
Could very well be - it's certainly incorrect, even though my local
installation didn't complain.
Anyway, I've uploaded
bobertuk wrote:
> Still the same. I noticed that
> C:\ProgramData\Squeezebox\Cache\InstalledPlugins\Plugins\LMSdbopt/install.xml
> contains 0.0.1
>
> Could that be the problem? Could the cache at your end contain the wrong
> version?
I just stopped LMS, edited 0.0.1 to
0.0.2, restarted LMS
bobertuk wrote:
> Still not working. Has downloaded into
> C:\ProgramData\Squeezebox\Cache\InstalledPlugins\Plugins folder but
> hasn't activated. Checkbox to install into plugins still showing so will
> go through cache cleanup and reboot again to see if that helps.
Still the same. I noticed
Roland0 wrote:
> My semi-automated deployment re-introduced the typo I fixed yesterday -
> sorry about that. Should work now.
Still not working. Has downloaded into
C:\ProgramData\Squeezebox\Cache\InstalledPlugins\Plugins folder but
hasn't activated. Checkbox to install into plugins still
bobertuk wrote:
> Have just tried updating to v0.0.2 and it fails to update. Have tried
> clearing web browser caches, uninstalling previous version and rebooting
> Windows 10 but no luck and now have no Database Optimizer at all. Any
> help appreciated.
My semi-automated deployment
mamema wrote:
> Sure, but how? I'm a very very poor Synology user (poor means shitty
> interface). As top or htop isn't included in the docker image
'gotop' (https://github.com/xxxserxxx/gotop/releases) or 'bottom'
(https://github.com/ClementTsang/bottom/releases) have binaries for most
Have just tried updating to v0.0.2 and it fails to update. Have tried
clearing web browser caches, uninstalling previous version and rebooting
Windows 10 but no luck and now have no Database Optimizer at all. Any
help appreciated.
2 x Touch
2 x Radio
2 x Boom
1 x Intel-NUC server/squeezelite
Roland0 wrote:
>
> Pretty confident that'll overwhelmingly be the new indices.
> However, this may actually be a scenario where multi-threading could be
> used, so mamema could check if more than one core is utilized the next
> time a trackstat refresh runs.
>
Sure, but how? I'm a very very
mherger wrote:
>
> From my poor memory I remember SQLite caches the index, not the data
> itself.
>
actually, it does cache the data:
SQLite docs wrote:
>
> maximum number of database disk pages that SQLite will hold in memory
>
mherger wrote:
>
> Therefore assigning a lot of data to
mherger wrote:
> > My, i admit rather short test reveals, that the trackscan scan task
> went
>
> I guess that should read "trackstat"?
>
Yes!
mamema's Profile: http://forums.slimdevices.com/member.php?userid=44662
My, i admit rather short test reveals, that the trackscan scan task went
I guess that should read "trackstat"?
don't from 3 hours to 4 min on LMS restart and from 3 hours to 15 min on
forced rescan. Even though modifying the SELECT statement to prevent
musicbrainz_id duplicates will bring the
mherger wrote:
> > Anything accessing the DB should benefit.
>
> That's rather general... where would I as a user see improvements? Not
> in a benchmark, but in LMS use.
> .
My, i admit rather short test reveals, that the trackscan scan task went
don't from 3 hours to 4 min on LMS restart
Anything accessing the DB should benefit.
That's rather general... where would I as a user see improvements? Not
in a benchmark, but in LMS use.
Eg. the scanner has totally different requirements than the server:
scanner is mostly writes, server is almost read-only. From my poor
memory I
Roland0 wrote:
> I think I've found the issue. Try removing the repository, restart, add
> it again, active the plugin, and restart (and report if there are any
> issues remaining). Thanks.
Yep, that's fixed it.
I've also checked that mmap is enabled it is:
$ sudo lsof |grep mem-r|grep
mherger wrote:
> > DB Optimizer is a plugin for LMS to improve it's database performance
> > and memory use.
> > More information can be found 'on the homepage'
> > (https://www.nexus0.net/pub/sw/lmsdboptim/).
>
> Would you have a rough idea what parts of LMS' functionality would take
>
>
mamema wrote:
>
> But as I said, you are doing more than index creating. I would be happy
> to test, unfortunately the plugin doesmt Show up after reboot.
It seems I initially had a typo in the repo.xml. However, at least on my
system, it now works.
If removing the repository, restarting,
Roland0 wrote:
> see previous comment for the rationale.
> _id=?)
You are doing more with your plugin than creating indexes, which might be
good, not tested yet.
But, dunno if it is an issue with pre 8.2 versions, because the indexes
were created fine within trackstat
DB Optimizer is a plugin for LMS to improve it's database performance
and memory use.
More information can be found 'on the homepage'
(https://www.nexus0.net/pub/sw/lmsdboptim/).
Would you have a rough idea what parts of LMS' functionality would take
advantage of which tweak? As Erland said it
Well, talk about a derail ;-)
So, to recap:
- the purpose of this plugin is to apply some tuning to the database
which should be generally beneficial (and is unrelated to TrackStat,
although TrackStat, using the DB heavily, may benefit as well)
- I added some indices to my database for
mamema wrote:
> which indexes do you mean? Except those from the temp table, we've
> discussed in the Dev Area, all needed indexes are there.
> Trackstat creates and deletes indexes and both is working.
see previous comment for the rationale. also:
Code:
explain query
Roland0 wrote:
>
> Which indices do you mean? The ones taken from TrackStat should use the
> same names afaik.
>
My mistake, I accidentally looked at the MySQL indexes TrackStat
creates. The SQLite indexes are named the same so no duplicates, sorry
for the confusion regarding this.
Erland
mamema wrote:
> which indexes do you mean? Except those from the temp table, we've
> discussed in the Dev Area, all needed indexes are there.
> Trackstat creates and deletes indexes and both is working.
I read the code a bit fast, the naming is the same so there arent any
duplicated indexes, I
erland wrote:
> Hopefully someone will take the time to update TrackStat so it creates
> the indexes by itself.
>
Would certainly be a better solution.
> The intention is that TrackStat should already today create some of the
> single column indexes but maybe that doesnÂt work since you have
erland wrote:
> Hopefully someone will take the time to update TrackStat so it creates
> the indexes by itself. The intention is that TrackStat should already
> today create some of the single column indexes but maybe that doesnt
> work .
which indexes do you mean? Except those from the temp
mherger wrote:
> >
> /var/lib/squeezeboxserver/cache/InstalledPlugins/Plugins/LMSdboptim/LMSdbopt/lib
>
> That's one folder too many. The plugin archive should extract without a
>
> sub-folder for Plugins.pm.
It extracts to LMSdbopt/Plugin.pm, which I think is correct. I think the
issue was
Hopefully someone will take the time to update TrackStat so it creates
the indexes by itself. The intention is that TrackStat should already
today create some of the single column indexes but maybe that doesnt
work since you have them in your plugin ? As its done now I think the
indexes will be
SpiderJon wrote:
> [[21-03-02 18:27:09.6876] main::init (388) Starting Logitech Media
> Server (v8.1.1, 1610364019, Thu Jan 14 06:24:07 CET 2021) perl 5.028001
> - arm-linux-gnueabihf-thread-multi-64int
> [21-03-02 18:27:11.4689] Slim::bootstrap::tryModuleLoad (286) Warning:
> Module
/var/lib/squeezeboxserver/cache/InstalledPlugins/Plugins/LMSdboptim/LMSdbopt/lib
That's one folder too many. The plugin archive should extract without a
sub-folder for Plugins.pm.
___
plugins mailing list
plugins@lists.slimdevices.com
Roland0 wrote:
> Anything in the server logs?
*[21-03-02 18:27:09.6876] main::init (388) Starting Logitech Media
Server (v8.1.1, 1610364019, Thu Jan 14 06:24:07 CET 2021) perl 5.028001
- arm-linux-gnueabihf-thread-multi-64int
[21-03-02 18:27:11.4689] Slim::bootstrap::tryModuleLoad (286)
SpiderJon wrote:
>
> However, I followed the installation routine
> but "DB Optimizer" didn't get installed.
>
> So I tried activating the plugin (ie, ticking the box), saving settings
> and restarting LMS three more times, but still no luck installing it
> it's just showing as an
Roland0 wrote:
> DB Optimizer is a plugin for LMS to improve its database performance and
> memory use.
> More information can be found 'on the homepage'
> (https://www.nexus0.net/pub/sw/lmsdboptim/).
Very interested in trying this.
However, I followed the installation routine
In the LMS
DB Optimizer is a plugin for LMS to improve it's database performance
and memory use.
More information can be found 'on the homepage'
(https://www.nexus0.net/pub/sw/lmsdboptim/).
'Various SW' (https://www.nexus0.net/pub/sw/): Web Interface | TUI |
Playlist Editor / Generator | Music
75 matches
Mail list logo