PDA

View Full Version : MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3



hunta
2005-04-26, 14:40
I've installed SlimServer 6.01 and SoftSqueeze 2.0b3 to get to grips with things while I wait for my SB2 to arrive.

This evening I tried integrating MusicMagic. It took a while to get the two talking to each other, solved only when the mysterious 'Use MusicMagic Playlist' radio button came to light.

As soon as that was checked things fell into place. However, slim.exe has been running at near 100% CPU ever since. I guess this is while it reads the MusicMagic data.

Is this normal, and if so how long can I expect it to last? At present the whole thing is unusable. I'm running XP SP2 on a machine with plenty of grunt.

Thanks in advance.

Andrew

kdf
2005-04-26, 14:56
Quoting hunta <hunta.1o484z (AT) no-mx (DOT) forums.slimdevices.com>:

> As soon as that was checked things fell into place. However, slim.exe
> has been running at near 100% CPU ever since. I guess this is while it
> reads the MusicMagic data.

it reads the data fairly fast. The problem is that MusicMagic itself can take
DAYS to run full profiles on all the songs. This means that the database is
always changing, so the server will keep reloading the musicmagic data 60's
seconds after it finishes the previous pass. If you go into server
settings->musicmagic, then you can set the MusicMagic Reload Interval to a very
large number. Alternatively, wait a while before turning 'use musicmagic' on
in server settings.

-kdf

tvoil
2005-04-26, 14:56
Make sure all your music is analyzed in Music Magic before you integrate it with Slim Server. Once it is it is best to do a wipe cache once you turn on Music Magic support in Slim Server.

Howard Darwen
2005-04-26, 14:58
Strange as MMM should be the process hogging CPE when it's scanning as far
as I can see. Am not sure what slim could be doing. I've actually turned
slim off while MMM is scanning my library - not because it was consuming
CPU, it's just MMM randomly stopped when I had slim on, and it seems to be
running more consistently now.

I have MMM scanning a subset of my files right now. Started on Sunday on a
batch of 2000 FLACS and it has about 400 to go. But it depends on whether
the songs in your collection already exist in the music magic online
database or not - or that's what I have assumed from what I have read. If a
track is already in the online database then MMM only seems to have to
generate a fingerprint (10 secs or so). If the track can't be found in the
database from the fingerprint then MMM does a detailed analysis - which
apparently takes 80% of the songs playtime. Most of my tunes seem not to be
there :)

Thx.

H.

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com]On Behalf Of hunta
Sent: 26 April 2005 22:40
To: discuss (AT) lists (DOT) slimdevices.com
Subject: [slim] MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3



I've installed SlimServer 6.01 and SoftSqueeze 2.0b3 to get to grips
with things while I wait for my SB2 to arrive.

This evening I tried integrating MusicMagic. It took a while to get the
two talking to each other, solved only when the mysterious 'Use
MusicMagic Playlist' radio button came to light.

As soon as that was checked things fell into place. However, slim.exe
has been running at near 100% CPU ever since. I guess this is while it
reads the MusicMagic data.

Is this normal, and if so how long can I expect it to last? At present
the whole thing is unusable. I'm running XP SP2 on a machine with
plenty of grunt.

Thanks in advance.

Andrew


--
hunta

hunta
2005-04-26, 15:10
Thanks for the feedback.

My library is fully analyzed, but I'll try turning analysis off in MMM and see whether that helps. One thing I wasn't sure of:

tvoil:

Make sure all your music is analyzed in Music Magic before you integrate it with Slim Server. Once it is it is best to do a wipe cache once you turn on Music Magic support in Slim Server

How do I do a wipe cache?

Thanks again.

kdf
2005-04-26, 15:46
> How do I do a wipe cache?

in the web browser, server settings->performance
click the Wipe Cache button. With 6.0.1, there have been some issues using wipe
cache, however. If you end up with strange results in the pagebar (the letters
at the top fo the browse lists), a server restart should clean them up.

-kdf

hunta
2005-04-26, 16:26
> How do I do a wipe cache?

in the web browser, server settings->performance
click the Wipe Cache button. With 6.0.1, there have been some issues using wipe
cache, however. If you end up with strange results in the pagebar (the letters
at the top fo the browse lists), a server restart should clean them up.

-kdf

Okay, I've tried all that and even uninstalled SS 6.01 and replaced it with the pre-release SS 6.02. Still exactly the same - slim.exe at near 100% if I enable MusicMagic integration. MM itself is doing nothing (CPU 0%, all songs analysed). The only way to break the deadlock seems to be to disable MM.

Any other ideas as to what I might be doing wrong?

Many thanks again.

Andrew

Dan Sully
2005-04-26, 16:31
* hunta shaped the electrons to say...

>Okay, I've tried all that and even uninstalled SS 6.01 and replaced it
>with the pre-release SS 6.02. Still exactly the same - slim.exe at near
>100% if I enable MusicMagic integration. MM itself is doing nothing (CPU
>0%, all songs analysed). The only way to break the deadlock seems to be
>to disable MM.

What version of MusicMagic are you running?

-D
--
vacation (n) : an extended trip away from home in search of inconvenient ways to connect to the Internet.

hunta
2005-04-27, 01:46
What version of MusicMagic are you running?

I downloaded it from the Predixis site a few days ago - 1.1.4, I believe.

ChrisB
2005-04-27, 02:35
I see this too, any time anything new gets added to the MM library. The
CPU usage does return to normal after a while, though.

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com] On Behalf Of hunta
Sent: 27 April 2005 00:26
To: discuss (AT) lists (DOT) slimdevices.com
Subject: [slim] Re: MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3



Okay, I've tried all that and even uninstalled SS 6.01 and replaced it
with the pre-release SS 6.02. Still exactly the same - slim.exe at near
100% if I enable MusicMagic integration. MM itself is doing nothing (CPU
0%, all songs analysed). The only way to break the deadlock seems to be
to disable MM.

Any other ideas as to what I might be doing wrong?

Many thanks again.

Andrew

hunta
2005-04-27, 06:05
I see this too, any time anything new gets added to the MM library. The
CPU usage does return to normal after a while, though.


Chris, how long is a while? I tried leaving it on the basis that it could take a while to initialise the first time but gave up after, say, an hour.

ChrisB
2005-04-27, 07:06
Sorry, I don't tend to keep a stopwatch on it, but it does seem to take
quite a while (30 mins?), even for small additions (10-20 tracks) to my
library.

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com] On Behalf Of hunta
Sent: 27 April 2005 14:05
To: discuss (AT) lists (DOT) slimdevices.com
Subject: [slim] Re: MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3


ChrisB Wrote:
> I see this too, any time anything new gets added to the MM library.
> The
> CPU usage does return to normal after a while, though.
>

Chris, how long is a while? I tried leaving it on the basis that it
could take a while to initialise the first time but gave up after, say,
an hour.


--
hunta

hunta
2005-04-27, 09:26
Sorry, I don't tend to keep a stopwatch on it, but it does seem to take
quite a while (30 mins?), even for small additions (10-20 tracks) to my
library.

Blimey, that's quite a hit. That means every time you add something to your library you need to wait for MM to analyse and verify it, then another half hour or whatever for SlimServer to 'catch up', with no tunes on the SB all the while.

Surely this isn't right?

Chris Howard
2005-04-27, 13:35
I have exactly the same issue and have tried a number of nightly builds (currently on 6.0.1 2781).

I actually bought MusicMagic specifically for the Slim integration, but given my machine grinds to a halt (AMD Athlon 64 3200+ with a GB of RAM) everytime I add tracks I have had to disable the MM -> Slim linkage.

I see 100% CPU for periods of around 30 minutes and this is not always related to adding new tracks - it seems to happen on a periodic basis - I am not sure if there is some refresh setting I have missed.

Dan Sully
2005-04-27, 13:40
* Chris Howard shaped the electrons to say...

>I actually bought MusicMagic specifically for the Slim integration, but
>given my machine grinds to a halt (AMD Athlon 64 3200+ with a GB of RAM)
>everytime I add tracks I have had to disable the MM -> Slim linkage.
>
>I see 100% CPU for periods of around 30 minutes and this is not always
>related to adding new tracks - it seems to happen on a periodic basis -
>I am not sure if there is some refresh setting I have missed.

Has MusicMagic completely finished analyzing your music?

-D
--
You have the puzzle pieces? Good, then turn off the damn walls.

Chris Howard
2005-04-27, 13:49
Yes - 100% completed - status bar is reading 0 / 0 / 0 so all tracks have been analysed / added etc.

Whether I start slim.exe manually or via the service it starts ok then within about 5-10 minutes it will ramp up to 100% CPU and stay there for 20-30 minutes. It will then come back down again. It will then jump back up again on regular intervals.

I saw a post sometime ago that it was related to Slim -> MM integration and so I disabled MM until I heard anymore. Seems others are still seeing the issue.

michael
2005-04-27, 14:16
kdf <slim-mail (AT) deane-freeman (DOT) com> writes:

> it reads the data fairly fast. The problem is that MusicMagic itself can take
> DAYS to run full profiles on all the songs. This means that the database is
> always changing, so the server will keep reloading the musicmagic data 60's
> seconds after it finishes the previous pass. If you go into server
> settings->musicmagic, then you can set the MusicMagic Reload Interval to a very
> large number. Alternatively, wait a while before turning 'use musicmagic' on
> in server settings.

I also recently filed a bug about this plugins effect on slimservers
cpu and memory usage. If this really is the reason for the cpu part
of that, then perhaps the default value should be significantly
larger?


-michael

Andy Bunn
2005-04-27, 14:44
> Yes - 100% completed - status bar is reading 0 / 0 / 0 so all tracks
> have been analysed / added etc.
>
> Whether I start slim.exe manually or via the service it starts ok then
> within about 5-10 minutes it will ramp up to 100% CPU and stay there
> for 20-30 minutes. It will then come back down again. It will then jump
> back up again on regular intervals.

I have the exact same issue. Running the current stable release of both
products and XP SP2.

-Andy

abunn
2005-04-27, 19:53
is that slimserver pegs the cpu occasionally at start-up when music magic is integrated (XP SP2). The attached image shows my desktop. slim.exe is the process pegging the cpu and music magic is done analyzing the library. I set the reload interval to 600 seconds but that didn't appear to help. If I kill music magic then the cpu returns to normal and slimserver functions normally. If I restart musicmagic, the slim process ramps up again and pegs the cpu. When things are frisky like this, if I just start music magic in the absence of slim, the cpu pegs and the MusicMagicMixer.exe is the culprit. But, no analysis is being done. At least, it claims not (there are a few tracks that mm lists as unanalyzable - too short). It's a drag; but clearly, music magic is to blame, not slim. Maybe this thread is happening in the wrong forum ;)

We need the kdf of mmm to save us.

-Andy

PS On the occasions that slim.exe doesn't peg the cpu continuously it pegs it every minute or so for ten seconds or so.

shaboyi
2005-04-27, 20:08
I also use MusicMagic intergrated with SlimServer and don't see this
behaviour once all the music is analyzed by MM and SlimServer has
fully scanned the library(prior to those two operations completing the
entire computer is maxed out, and I typically stop the slimserver
service, let MM do its thing, and then start slimserver again). This
is true of a couple of Windows machines (W2K, XP, varying processors
from PIII 933 to PIV 2.8, 512 MB RAM and above)

Once complete, I see no unsual cpu usage. But I also do not
regularly add music. But if this is the main issue, why not add the
music before you go to bed?


On 4/27/05, abunn <abunn.1o6h5n (AT) no-mx (DOT) forums.slimdevices.com> wrote:
>
> is that slimserver pegs the cpu occasionally at start-up when music
> magic is integrated (XP SP2). The attached image shows my desktop.
> slim.exe is the process pegging the cpu and music magic is done
> analyzing the library. I set the reload interval to 600 seconds but
> that didn't appear to help. If I kill music magic then the cpu returns
> to normal and slimserver functions normally. If I restart musicmagic,
> the slim process ramps up again and pegs the cpu. When things are
> frisky like this, if I just start music magic in the absence of slim,
> the cpu pegs and the MusicMagicMixer.exe is the culprit. But, no
> analysis is being done. At least, it claims not (there are a few tracks
> that mm lists as unanalyzable - too short). It's a drag; but clearly,
> music magic is to blame, not slim. Maybe this thread is happening in
> the wrong forum ;)
>
> We need the kdf of mmm to save us.
>
> -Andy
>
> PS On the occasions that slim.exe doesn't peg the cpu continuously it
> pegs it every minute or so for ten seconds or so.
>
> +-------------------------------------------------------------------+
> |Filename: new-1.jpg |
> |Download: http://forums.slimdevices.com/vb/attachment.php?attachmentid=103|
> +-------------------------------------------------------------------+
>
> --
> abunn
>

Wendell Hicken
2005-04-27, 21:25
On 4/27/05, abunn <abunn.1o6h5n (AT) no-mx (DOT) forums.slimdevices.com> wrote:
> If I just start music magic in the absence of slim,
> the cpu pegs and the MusicMagicMixer.exe is the culprit. But, no
> analysis is being done. At least, it claims not (there are a few tracks
> that mm lists as unanalyzable - too short). It's a drag; but clearly,
> music magic is to blame, not slim. Maybe this thread is happening in
> the wrong forum ;)

Are you using Watch folders? Maybe MMM is scanning a large/remote
disk looking for new stuff?

Wendell (MMM Developer)

kdf
2005-04-27, 23:42
On 27-Apr-05, at 9:25 PM, Wendell Hicken wrote:

> On 4/27/05, abunn <abunn.1o6h5n (AT) no-mx (DOT) forums.slimdevices.com> wrote:
>> If I just start music magic in the absence of slim,
>> the cpu pegs and the MusicMagicMixer.exe is the culprit. But, no
>> analysis is being done. At least, it claims not (there are a few
>> tracks
>> that mm lists as unanalyzable - too short). It's a drag; but clearly,
>> music magic is to blame, not slim. Maybe this thread is happening in
>> the wrong forum ;)
>
> Are you using Watch folders? Maybe MMM is scanning a large/remote
> disk looking for new stuff?
>
> Wendell (MMM Developer)

aha! the kdf of mmm, or am I the whicken of slim?? ;)

-k

hunta
2005-04-28, 01:24
Are you using Watch folders? Maybe MMM is scanning a large/remote
disk looking for new stuff?

Wendell (MMM Developer)

That could be it, I guess - I am using Watch Folders. I'll try turning them off in MM and seeing if that helps. I'll also post a similar thread on the Predexis forum.

relen
2005-04-28, 02:55
I did a series of test lookups on their web site search front-end and it is obviously only checking a couple of locations carrying mainstream current albums. Is this the same on-line database as their product uses, or is the latter something more sophisticated?

The web-based search found under 15% of the albums or artists I asked it about (deliberately awkward ones of course): Album Art Aggregator is better at locating album art than that, using very similar sources.

I guess the problem is that my collection includes things like Sergey Kuryokhin's 'Sparrow Oratorium'and Huub de Lange's 'Life and death in a street organ'...(both artists are in the database at AllMusic.com, I wonder if anyone's thought of using them for track/artwork lookups?)

All this presumably means that MusicMagic would have to do complete analysis on a significant proportion of my collection. What's the general view on this product, especially as one evidently has to subscribe to get Slimserver integration? Is it worth the effort?

Thanks,
--Richard E

abunn
2005-04-28, 05:32
> What's the
> general view on this product, especially as one evidently has to
> subscribe to get Slimserver integration? Is it worth the effort?

It's great (when it works I guess - given the nature of this thread). I have
a 10k song library with seriously eclectic spread (I suppose that in 10k
tracks it's easy to be eclectic). Almost every time I make a mix I listen to
something of quality that I haven't heard in years...I almost always think,
"How did it know?"

It's a great product and once the integration is solved then it'll be near
perfect. I do wonder how it works. I imagine through some type of eigen
solution as it's so fast. Does anybody know?

Patrick Dixon
2005-04-28, 07:00
I guess the problem is that my collection includes things like Sergey Kuryokhin's 'Sparrow Oratorium'and Huub de Lange's 'Life and death in a street organ'...(both artists are in the database at AllMusic.com, I wonder if anyone's thought of using them for track/artwork lookups?)
Music Collector uses AllMusic - so I assume their MP3 variant does too.

Jim Forsythe
2005-04-28, 07:36
I've been having the same problem. If I turn MMM on, and integrate it
with slimserver, within a single song play, slimserver "locks up". The
web interface won't respond, and after the buffer empties, it stops
playing. This is on Linux (red hat 9).

I don't see an option for watch folders anywhere on MMM - where is it?

Jim

> -----Original Message-----
> From: hunta [mailto:hunta.1o6wfn (AT) no-mx (DOT) forums.slimdevices.com]
> Sent: Thursday, April 28, 2005 4:25 AM
> To: discuss (AT) lists (DOT) slimdevices.com
> Subject: [slim] Re: MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3
>
>
>
> Wendell Hicken Wrote:
> >
> > Are you using Watch folders? Maybe MMM is scanning a large/remote
> > disk looking for new stuff?
> >
> > Wendell (MMM Developer)
>
> That could be it, I guess - I am using Watch Folders. I'll
> try turning them off in MM and seeing if that helps. I'll
> also post a similar thread on the Predexis forum.
>
>
> --
> hunta
>
>

hunta
2005-04-28, 07:47
I don't see an option for watch folders anywhere on MMM - where is it?

The second half of this page should give you what you need to know:
http://music.predixis.com/playlistadj.html

Wendell Hicken
2005-04-28, 10:50
On 4/28/05, relen <relen.1o70tz (AT) no-mx (DOT) forums.slimdevices.com> wrote:
> I did a series of test lookups on their web site search front-end and it
> is obviously only checking a couple of locations carrying mainstream
> current albums. Is this the same on-line database as their product
> uses, or is the latter something more sophisticated?

I think you mean the cdideas.predixis.com website - if so, this is only using
a small subset of the whole database which has been connected to Amazon and
CD Baby. The whole database is over 5,000,000 songs right now.

Wendell

Wendell Hicken
2005-04-28, 10:51
On 4/28/05, Jim Forsythe <jimforsythe (AT) metrocast (DOT) net> wrote:
> I've been having the same problem. If I turn MMM on, and integrate it
> with slimserver, within a single song play, slimserver "locks up". The
> web interface won't respond, and after the buffer empties, it stops
> playing. This is on Linux (red hat 9).
>
> I don't see an option for watch folders anywhere on MMM - where is it?

We don't have watch folders on Linux yet, so this is likely a different issue
(if watch folders are indeed the cause of the other issue).

Wendell

Wendell Hicken
2005-04-28, 10:54
On 4/28/05, relen <relen.1o70tz (AT) no-mx (DOT) forums.slimdevices.com> wrote:
> All this presumably means that MusicMagic would have to do complete
> analysis on a significant proportion of my collection. What's the
> general view on this product, especially as one evidently has to
> subscribe to get Slimserver integration? Is it worth the effort?

As far as the subscription goes, I can get you free access until your collection
is up and running, and you see how you like it for yourself. No need to spend
any money until you're sure you are happy with the results. You can either
use the "Trial Key" button in version 1.1.5, or email me directly.

Wendell

Patrick Dixon
2005-04-28, 11:28
This is probably a dumb question but I can I use MM on a Windows machine with SS running on a Linux server (and have them talk to each other)?

hunta
2005-04-28, 11:50
I've tried everything I can think of now, including turning off Watch Folders, and there's no change in the situation. I left it grinding away for about 45 minutes, during which time my system was unusable, before giving up and closing down MMM.

This is a real shame as I love the concept but can't make it work. I guess I could try MoodLogic, but having looked at both initially MMM seems easier and I felt gave better results.

If anyone has any other thoughts I'd love to hear them.

Thanks,

Andrew

abunn
2005-04-28, 12:02
> guess I could try MoodLogic,

I did and wasn't thrilled with it. MMM is a better music mixer. Do you have
some fraction of unanalyzable tracks? I do. That is, I have no songs which
have not yet been analyzed but I do have songs which are unanalyzable. Can
that be part of it?

whicken (the kdf of MMM) will save us!

hunta
2005-04-28, 12:18
I do have unanalyzable tracks, though all activity which can be completed has been. I don't know if it's the unanalyzable tracks which are causing the problem though as Task Manager is reporting that it's slim.exe which is hogging the CPU. I'd have thought it would be MMM hogging CPU if it's something to do with analyzing.

abunn
2005-04-28, 12:26
Right, but when I have this situation and I shutdown both slim and
musicmagic and restart musicmagic only, then musicmagic hogs the cpu. this
is leading me to believe that musicmagic is the culprit.

> -----Original Message-----
> From: discuss-bounces (AT) lists (DOT) slimdevices.com
> [mailto:discuss-bounces (AT) lists (DOT) slimdevices.com]On Behalf Of hunta
> Sent: Thursday, April 28, 2005 3:19 PM
> To: discuss (AT) lists (DOT) slimdevices.com
> Subject: [slim] Re: MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3
>
>
>
> I do have unanalyzable tracks, though all activity which can be
> completed has been. I don't know if it's the unanalyzable tracks which
> are causing the problem though as Task Manager is reporting that it's
> slim.exe which is hogging the CPU. I'd have thought it would be MMM
> hogging CPU if it's something to do with analyzing.
>
>
> --
> hunta
>

hunta
2005-04-29, 02:05
After more experimentation ('testing' would be too grand a word!) last night, I found that there was some progress, it's just extremely slow. I left my machine running MMM and SlimServer all evening, probably around 3 hours, and of 2,000 songs I had the 'mm' icon down to early 'C' entries. After another hour, most of the 'C's had been done.

I don't know the precise alphabetical distribution of my collection, but this implies it's got a way to go yet. I've left it running today while I'm in the office. Hopefully it'll reach a point where SlimServer knows all about the MMM library and will then, with a sigh of relief, slip back into normal running.

Does this fit with other people's experience when first introducing MMM and SS to each other? People have posted issues of 30 min delays when adding an album's worth of songs, but nothing about the initial scan. Maybe it's something with these particular versions - if you initially scanned your collection with earlier versions you're okay?

That leads to another question. If I upgrade either MMM or SS, will I have to repeat this painful process?

Any help would be really appreciated.

Thanks,

Andrew

Howard Darwen
2005-04-29, 07:05
from what i've worked out, it MMM does something like this ...

- first, MMM generates a fingerprint for each song, taking about 10 secs per
song
- if MMM can find that song from the fingerprint in the online database then
it can get the full analysis data from there (cos someone else has already
done it for you and it's been uploaded to the online database!)
- if the song is not in the online database then it takes roughly 80% of the
playtime of the song to analyse

for me it took 3 to 4 days continuous running to analyse 2000 FLACs. it's
hard to determine, but I think about half my music was not on the server.

on the first pass, i switched slim off for the analysis. i now have it on
again, but the analysis process does not work so well when both are on.
which is a total pain in the arse as i add music all the time and i don't
want to keep switching slim off.

if slim is on when you're analysing, from my experience it will either slow
it down massively or just grind to a halt completely. it also indicates slim
is hogging the CPU rather than MMM (which seems strange).

so i'm thinking of an alternative strategy where i do the analysis on my
normal PC, archive the info into the tags, then upload to the server. which
means the data is already there and MMM won't have to analyse. i don't
really like this though, and it would be better if things worked with both
services on at the server.

thx.

h.

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com]On Behalf Of hunta
Sent: 29 April 2005 10:06
To: discuss (AT) lists (DOT) slimdevices.com
Subject: [slim] Re: MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3



After more experimentation ('testing' would be too grand a word!) last
night, I found that there was some progress, it's just extremely slow.
I left my machine running MMM and SlimServer all evening, probably
around 3 hours, and of 2,000 songs I had the 'mm' icon down to early
'C' entries. After another hour, most of the 'C's had been done.

I don't know the precise alphabetical distribution of my collection,
but this implies it's got a way to go yet. I've left it running today
while I'm in the office. Hopefully it'll reach a point where SlimServer
knows all about the MMM library and will then, with a sigh of relief,
slip back into normal running.

Does this fit with other people's experience when first introducing MMM
and SS to each other? People have posted issues of 30 min delays when
adding an album's worth of songs, but nothing about the initial scan.
Maybe it's something with these particular versions - if you initially
scanned your collection with earlier versions you're okay?

That leads to another question. If I upgrade either MMM or SS, will I
have to repeat this painful process?

Any help would be really appreciated.

Thanks,

Andrew


--
hunta

Howard Darwen
2005-04-29, 07:12
by the way, i just upgraded MMM and it picks up all the analysis data
without a problem. so you don't need to fret about upgrades.

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com]On Behalf Of Howard Darwen
Sent: 29 April 2005 15:05
To: Slim Devices Discussion
Subject: RE: [slim] Re: MusicMagic, SlimServer 6.01 and SoftSqueeze
2.0b3


from what i've worked out, it MMM does something like this ...

- first, MMM generates a fingerprint for each song, taking about 10 secs per
song
- if MMM can find that song from the fingerprint in the online database then
it can get the full analysis data from there (cos someone else has already
done it for you and it's been uploaded to the online database!)
- if the song is not in the online database then it takes roughly 80% of the
playtime of the song to analyse

for me it took 3 to 4 days continuous running to analyse 2000 FLACs. it's
hard to determine, but I think about half my music was not on the server.

on the first pass, i switched slim off for the analysis. i now have it on
again, but the analysis process does not work so well when both are on.
which is a total pain in the arse as i add music all the time and i don't
want to keep switching slim off.

if slim is on when you're analysing, from my experience it will either slow
it down massively or just grind to a halt completely. it also indicates slim
is hogging the CPU rather than MMM (which seems strange).

so i'm thinking of an alternative strategy where i do the analysis on my
normal PC, archive the info into the tags, then upload to the server. which
means the data is already there and MMM won't have to analyse. i don't
really like this though, and it would be better if things worked with both
services on at the server.

thx.

h.

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com]On Behalf Of hunta
Sent: 29 April 2005 10:06
To: discuss (AT) lists (DOT) slimdevices.com
Subject: [slim] Re: MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3



After more experimentation ('testing' would be too grand a word!) last
night, I found that there was some progress, it's just extremely slow.
I left my machine running MMM and SlimServer all evening, probably
around 3 hours, and of 2,000 songs I had the 'mm' icon down to early
'C' entries. After another hour, most of the 'C's had been done.

I don't know the precise alphabetical distribution of my collection,
but this implies it's got a way to go yet. I've left it running today
while I'm in the office. Hopefully it'll reach a point where SlimServer
knows all about the MMM library and will then, with a sigh of relief,
slip back into normal running.

Does this fit with other people's experience when first introducing MMM
and SS to each other? People have posted issues of 30 min delays when
adding an album's worth of songs, but nothing about the initial scan.
Maybe it's something with these particular versions - if you initially
scanned your collection with earlier versions you're okay?

That leads to another question. If I upgrade either MMM or SS, will I
have to repeat this painful process?

Any help would be really appreciated.

Thanks,

Andrew


--
hunta

Ert
2005-04-29, 08:26
Hiya.

On Apr 29, 2005, at 5:05 AM, hunta wrote:
> After more experimentation ('testing' would be too grand a word!) last
> night, I found that there was some progress, it's just extremely slow.
[...]
> Does this fit with other people's experience when first introducing MMM
> and SS to each other? People have posted issues of 30 min delays when
> adding an album's worth of songs, but nothing about the initial scan.
> Maybe it's something with these particular versions - if you initially
> scanned your collection with earlier versions you're okay?

I can't speak for MM + SS together, but my MM alone has been working on
my 21,000 song library for weeks now. The whole process was going to
take about 300 hours on my 1 GHz G4 Powerbook, and I leave it working
overnights when I remember.

Thankfully, it has long since finished finding candidate identities for
all the songs and they're available for matches, but the final
verification fingerprinting is quite slow.

> That leads to another question. If I upgrade either MMM or SS, will I
> have to repeat this painful process?

Nope, I've upgraded through a couple of beta versions and the library
has stayed intact.

- Ert

hunta
2005-04-29, 11:01
Well, having left MMM and SS thrashing away while I went to the office, I got back to find that they've sorted themselves out and all songs have the 'mm' logo next to them. CPU is down to standard, though it does seem to have the odd busy moment which also causes a pause in playing. My SB2 hasn't arrived yet, so I'm just running SoftSqueeze - hopefull a hardware buffer will take care of that.

So, all in all it seems to work. Just takes a healthy dose of patience... I guess the reason why SS takes so long to read a fully analyzed and verified MMM library will remain a mystery for now.

Thanks to all for you input.

Andrew

abunn
2005-04-29, 11:15
> So, all in all it seems to work. Just takes a healthy dose of
> patience... I guess the reason why SS takes so long to read a fully
> analyzed and verified MMM library will remain a mystery for now.

There's a bug report on this and hopfully it will be fixed soon

http://bugs.slimdevices.com/show_bug.cgi?id=1446

Howard Darwen
2005-04-29, 11:16
i think it's the analysis that takes the time. and having slim up while your
analysing doesn't seem a good idea. once it's all done though, it is really
kool.

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com]On Behalf Of hunta
Sent: 29 April 2005 19:02
To: discuss (AT) lists (DOT) slimdevices.com
Subject: [slim] Re: MusicMagic, SlimServer 6.01 and SoftSqueeze 2.0b3



Well, having left MMM and SS thrashing away while I went to the office,
I got back to find that they've sorted themselves out and all songs
have the 'mm' logo next to them. CPU is down to standard, though it
does seem to have the odd busy moment which also causes a pause in
playing. My SB2 hasn't arrived yet, so I'm just running SoftSqueeze -
hopefull a hardware buffer will take care of that.

So, all in all it seems to work. Just takes a healthy dose of
patience... I guess the reason why SS takes so long to read a fully
analyzed and verified MMM library will remain a mystery for now.

Thanks to all for you input.

Andrew


--
hunta

hunta
2005-04-29, 11:18
Analysis was fully complete before SS came on the scene.

I spoke too soon. I'm very disappointed to find that, having closed and restarted MMM and SS CPU has become an issue again. I'm starting to run out of ideas (other than give up on MMM).

hunta
2005-04-29, 11:35
>
There's a bug report on this and hopfully it will be fixed soon

http://bugs.slimdevices.com/show_bug.cgi?id=1446
Interesting - thanks for pointing this out.

I'm not sure this is exactly the same however, though I'd imagine they're related. The logged issue talks about a gradual degradation in performance and is triggered by the SS's reload of the MMM library. In the scenario I'm experiencing, I can start MMM & SS, everything's fine for maybe 60 secs, then SS maxes out on the CPU. My reload interval is set to 3600 seconds. To verify this I set the reload interval to 0 (don't reload), as suggested in the bug detail, and it had no effect.

Could someone from Slim advise on the best way for me to contribute this? It might be useful info on bug 1446, or a different issue. I'm happy to run traces etc. if it would help.

Thanks.

abunn
2005-04-29, 12:07
> the scenario I'm experiencing, I can start MMM & SS, everything's fine
> for maybe 60 secs, then SS maxes out on the CPU. My reload interval is
> set to 3600 seconds. To verify this I set the reload interval to 0
> (don't reload), as suggested in the bug detail, and it had no effect.

Ditto. I advise patience.

hunta
2005-04-29, 12:28
During the 60 sec interval between starting SS and everything grinding to a halt I managed to turn on d_musicmagic and got this output (I've trimmed the path & file names). Not sure whether this means SS or MMM is re-evaluating the songs, but something definitely thinks it needs it even though nothing's altered since it last completed:

=============

2005-04-29 20:12:32.7343 MusicMagic: music library has changed!
2005-04-29 20:12:32.7344 MusicMagic: using musicmagic: Predixis MusicMagic Mixer. Version 1.1.4

2005-04-29 20:12:32.7345 MusicMagic: start export
2005-04-29 20:12:32.7350 MusicMagic: export mode is: start
2005-04-29 20:12:37.7346 MusicMagic: Got 2054 song(s).
2005-04-29 20:12:37.7348 MusicMagic: export mode is: songs
2005-04-29 20:12:47.7343 MusicMagic: Exporting song 1: C:\Documents and Settings\Andrew\My Docum
2005-04-29 20:12:47.7810 MusicMagic: export mode is: songs
2005-04-29 20:12:52.7351 MusicMagic: Exporting song 2: C:\Documents and Settings\Andrew\My Docum
2005-04-29 20:12:52.7475 MusicMagic: export mode is: songs
2005-04-29 20:13:03.7509 MusicMagic: Exporting song 3: C:\Documents and Settings\Andrew\My Docum
2005-04-29 20:13:03.7613 MusicMagic: export mode is: songs
2005-04-29 20:13:03.8101 MusicMagic: Exporting song 4: C:\Documents and Settings\Andrew\My Docum
2005-04-29 20:13:03.8207 MusicMagic: export mode is: songs
2005-04-29 20:13:04.1285 MusicMagic: Exporting song 5: C:\Documents and Settings\Andrew\My Docum
2005-04-29 20:13:04.1393 MusicMagic: export mode is: songs
2005-04-29 20:13:06.9849 MusicMagic: Exporting song 6: C:\Documents and Settings\Andrew\My Docum
2005-04-29 20:13:07.9982 MusicMagic: export mode is: songs
2005-04-29 20:13:16.7355 MusicMagic: Exporting song 7: C:\Documents and Settings\Andrew\My Docum

===============

kdf
2005-04-29, 12:40
Quoting hunta <hunta.1o9inb (AT) no-mx (DOT) forums.slimdevices.com>:

>
> Analysis was fully complete before SS came on the scene.
>
> I spoke too soon. I'm very disappointed to find that, having closed and
> restarted MMM and SS CPU has become an issue again. I'm starting to run
> out of ideas (other than give up on MMM).
>
please see the comments from bug 1446, and consider providing similar info to
what has been requested there.

We all want to see everyone happy and all problems fixed, but frankly, it gets
easier to simply choose to ignore as more and more people stomp up and down
saying "its broken", when we can't see it for ourselves (I know I've unfairly
appeared to single you out here, but this is a public forum so realise this is
a more global statement ).

If you are certain that 1446 is not the problem you have, then more information
is needed. I'd love to be able to tell you exactly what to look for, but that
would require that I have the slightest idea of what is causing your problem,
without it being even remotely related to 1446. I have tried to say things
like this before, and get barked at. The simple fact is that with all the
noise, you have to distinguish yourself in order to get more people on the job.
SilmDevices are working on every problem, but they are a limited resource.
The software is open, so anyone can help, but they are not bound into helping.
In some sense, it has to be easy to help.

I would start by running command line, so that you can see everything that is
going on with slimserver in real time. I'd start with d_musicmagic and maybe
d_info. If its not a constant rescan that is bogging things down, then you
will see very little activity from this. If it does start scanning, despite 0
reload interval or 3600 interval, then that it concrete info that may be
useful. If nothing shows up, you need to try d_remotestream, or d_http perhaps.
100% CPU is a very imprecise symptom, and we to eliminate or isolate the
primary.Otherwise, its like trying to navigate a minefield by throwing shotput
:)

If I've missed your logs, I apologise. It is very hard to keep up with the
volume. Please resubmit them to bugszilla for safe keeping.

-kdf

kdf
2005-04-29, 12:44
Quoting hunta <hunta.1o9lvz (AT) no-mx (DOT) forums.slimdevices.com>:

>
> During the 60 sec interval between starting SS and everything grinding
> to a halt I managed to turn on d_musicmagic and got this output (I've
> trimmed the path & file names). Not sure whether this means SS or MMM
> is re-evaluating the songs, but something definitely thinks it needs it
> even though nothing's altered since it last completed:

ah, now we're getting somewhere...

try adding d_remotestream to this, I think it should show the calls to the MMM
api via HTTP. I can't recall offhand if we are showing the responses, but if
not we'll have to add something to confirm. What does MMM show for numbers at
the very bottom of its window? If you pulldown the library menu, does it say
'start analysis' or 'stop analysis'. There was a component that seemed to take
more than a week when I first used MMM before slimserver stopped seeing a
changed timestamp every time it checked.

-kdf

hunta
2005-04-29, 13:43
kdf, I apologise if my requests for help have sounded like whingeing - I never intended that. Like many others on this forum I'm sure, I work in development myself and know very well the 'It doesn't work' user feedback. That said, I've not been using SS very long, so much of what I'm requesting is how to go about providing you guys with information that will help. The debug logs are something I discovered how to use this evening. You also mention trying to run from the command line - I'm afraid I'll need a bit more info to give this a go.

To respond to your specific points:

Here's a snippet from a dump with the d_remotestream as well as the d_musicmagic flags set (it gets very repetitive after this):

=================

2005-04-29 21:09:46.9064 Opening connection to http://localhost:10002/api/cacheid: [localhost on port 10002 with path /api/cacheid with timeout 5]
2005-04-29 21:09:46.9085 Request: GET /api/cacheid HTTP/1.0
Host: localhost:10002
User-Agent: iTunes/3.0 (MSWin32; SlimServer 6.0.2)
Accept: */*
Cache-Control: no-cache
Connection: close
Icy-MetaData:1

2005-04-29 21:09:46.9093 Response: HTTP/1.0 200 OK
2005-04-29 21:09:46.9095 header: Content-Type: text/plain
2005-04-29 21:09:46.9097 header:
2005-04-29 21:09:46.9097 Recieved final blank line...
2005-04-29 21:09:46.9097 opened stream!
2005-04-29 21:09:53.9062 MusicMagic: music library has changed!
2005-04-29 21:09:53.9064 MusicMagic: using musicmagic: Predixis MusicMagic Mixer. Version 1.1.4

2005-04-29 21:09:53.9064 MusicMagic: start export
2005-04-29 21:09:53.9068 MusicMagic: export mode is: start
2005-04-29 21:09:53.9069 Opening connection to http://localhost:10002/api/getSongCount: [localhost on port 10002 with path /api/getSongCount with timeout 5]
2005-04-29 21:09:53.9143 Request: GET /api/getSongCount HTTP/1.0
Host: localhost:10002
User-Agent: iTunes/3.0 (MSWin32; SlimServer 6.0.2)
Accept: */*
Cache-Control: no-cache
Connection: close
Icy-MetaData:1

2005-04-29 21:09:53.9155 Response: HTTP/1.0 200 OK
2005-04-29 21:09:53.9158 header: Content-Type: text/plain
2005-04-29 21:09:53.9159 header:
2005-04-29 21:09:53.9160 Recieved final blank line...
2005-04-29 21:09:53.9160 opened stream!
2005-04-29 21:10:00.4047 MusicMagic: Got 2054 song(s).
2005-04-29 21:10:00.4048 MusicMagic: export mode is: songs
2005-04-29 21:10:00.4049 Opening connection to http://localhost:10002/api/getSong?index=0: [localhost on port 10002 with path /api/getSong?index=0 with timeout 5]
2005-04-29 21:10:00.4062 Request: GET /api/getSong?index=0 HTTP/1.0
Host: localhost:10002
User-Agent: iTunes/3.0 (MSWin32; SlimServer 6.0.2)
Accept: */*
Cache-Control: no-cache
Connection: close
Icy-MetaData:1

2005-04-29 21:10:00.4337 Response: HTTP/1.0 200 OK
2005-04-29 21:10:00.4340 header: Content-Type: text/plain
2005-04-29 21:10:00.4341 header:
2005-04-29 21:10:00.4341 Recieved final blank line...
2005-04-29 21:10:00.4342 opened stream!
2005-04-29 21:10:05.9069 MusicMagic: Exporting song 1: C:\Documents and Settings\Andrew\My Documents\My Music\iTunes\iTunes Music\Belly\Star\01 Someone To Die For.mp3
2005-04-29 21:10:05.9480 MusicMagic: export mode is: songs
2005-04-29 21:10:05.9481 Opening connection to http://localhost:10002/api/getSong?index=1: [localhost on port 10002 with path /api/getSong?index=1 with timeout 5]
2005-04-29 21:10:06.1826 Request: GET /api/getSong?index=1 HTTP/1.0
Host: localhost:10002
User-Agent: iTunes/3.0 (MSWin32; SlimServer 6.0.2)
Accept: */*
Cache-Control: no-cache
Connection: close
Icy-MetaData:1

2005-04-29 21:10:06.2135 Response: HTTP/1.0 200 OK
2005-04-29 21:10:06.2138 header: Content-Type: text/plain
2005-04-29 21:10:06.2139 header:
2005-04-29 21:10:06.2139 Recieved final blank line...
2005-04-29 21:10:06.2140 opened stream!
2005-04-29 21:10:06.2206 MusicMagic: Exporting song 2: C:\Documents and Settings\Andrew\My Documents\My Music\iTunes\iTunes Music\Belly\Star\02 Angel.mp3

==============

Also:
> What does MMM show for numbers at the very bottom of its window?
Songs: 2,054 (2,054 ready)

> If you pulldown the library menu, does it say 'start analysis' or 'stop analysis'
It says 'Start analysis'

Hopefully this is useful info.

If you think, as I suspect, that all this is more suited to an entry in Bugzilla rather than here on the forum, either as additional info on an existing issue or as a new issue, please let me know. I'm more than happy to put this through the proper channels and provide as much information as possible with a view to enabling others to recreate and ultimately help resolve the problem.

Many thanks for your help so far.

Andrew

Spies
2005-04-29, 13:57
Version 1.1.5 of MusicMagic Mixer is out now. I have no idea if it
will help you, but I figure it is worth a try.

> 2005-04-29 21:09:53.9064 MusicMagic: using musicmagic: Predixis
> MusicMagic Mixer. Version 1.1.4

abunn
2005-04-29, 14:13
> Version 1.1.5 of MusicMagic Mixer is out now. I have no idea if it
> will help you, but I figure it is worth a try.

I'm using 1.1.5 with the same problem. I'll, too, stop whingeing, and post
some logs to the bug reports.
-Andy

kdf
2005-04-29, 14:38
Quoting hunta <hunta.1o9pcz (AT) no-mx (DOT) forums.slimdevices.com>:

>
> kdf, I apologise if my requests for help have sounded like whingeing - I
> never intended that. Like many others on this forum I'm sure, I work in
> development myself and know very well the 'It doesn't work' user
> feedback. That said, I've not been using SS very long, so much of what
> I'm requesting is how to go about providing you guys with information
> that will help. The debug logs are something I discovered how to use
> this evening. You also mention trying to run from the command line -
> I'm afraid I'll need a bit more info to give this a go.

Youhave certainly been nowhere near the worst offender, so dont take it as a
direct attack on you. I felt, however, that it needed to be said as a general
point to anyone reading these or taking part in teh 'me too' camp. Logs are a
geniune need unless the bug is clearly repeatable.


> To respond to your specific points:
>
> Here's a snippet from a dump with the d_remotestream as well as the
> d_musicmagic flags set (it gets very repetitive after this):


somethign very odd, given that your reponses indicate that anaylis shoudln't be
running. I was hoping to see more of the responses from MMM in the log, but I
guess I've forgotten amongst all the debug that I had in there when i was
reworking it. I'll add a bit more in there hopefully for tomorrow's nightly
build, to track the scan trigger and we'll have to see the logs again.

In the meantime, it might be worth trying out MMM 1.1.5, which I believe has
been released a couple days ago. That is what I've been using in the recent
tests.
=kdf

kdf
2005-04-29, 14:49
Quoting hunta <hunta.1o9pcz (AT) no-mx (DOT) forums.slimdevices.com>:

> 2005-04-29 21:09:53.9062 MusicMagic: music library has changed!
> 2005-04-29 21:09:53.9064 MusicMagic: using musicmagic: Predixis
> MusicMagic Mixer. Version 1.1.4
>
> 2005-04-29 21:09:53.9064 MusicMagic: start export
> 2005-04-29 21:09:53.9068 MusicMagic: export mode is: start


I have some more questions about this part, for you or anyone seeing the same
thing in their logs.

Is this at startup?

There should be mention of the reload interval after the mention that the
library changed. Have you managed to actually complete a scan before?

-kdf

kdf
2005-04-29, 14:49
Quoting hunta <hunta.1o9pcz (AT) no-mx (DOT) forums.slimdevices.com>:

> 2005-04-29 21:09:53.9062 MusicMagic: music library has changed!
> 2005-04-29 21:09:53.9064 MusicMagic: using musicmagic: Predixis
> MusicMagic Mixer. Version 1.1.4
>
> 2005-04-29 21:09:53.9064 MusicMagic: start export
> 2005-04-29 21:09:53.9068 MusicMagic: export mode is: start


I have some more questions about this part, for you or anyone seeing the same
thing in their logs.

Is this at startup?

There should be mention of the reload interval after the mention that the
library changed. Have you managed to actually complete a scan before?

-kdf

hunta
2005-04-29, 15:03
I'll add a bit more in there hopefully for tomorrow's nightly build, to track the scan trigger and we'll have to see the logs again.

That sounds good. Unfortunately I'll be away for a week, but it looks like abunn is experiencing the same issue so may be able to provide feedback (assuming his logs look the same to start with). Otherwise I'll be back on to it next Monday.


In the meantime, it might be worth trying out MMM 1.1.5, which I believe has been released a couple days ago. That is what I've been using in the recent tests.

I've upgraded to 1.1.5 and same story - as seen also by abunn.


Is this at startup?

Yes, that's the beginning of the log onwards. I wasn't able to get the debug settings to save between sessions though, so I don't know what SS is doing on initialise (i.e. between start-up and me enabling the debug).


Have you managed to actually complete a scan before?

I assume you mean SS scanning the MMM library. Yes - I left it running all day today while I was at work, and it had finished when I got back. Everything worked basically fine (a few CPU peaks causing minor playback interruptions on the SoftSqueeze) until I closed MMM & SS down and restarted them, at which point it seemed to have lost all knowledge and it was back to square 1.

kdf
2005-04-29, 15:11
Quoting hunta <hunta.1o9t2b (AT) no-mx (DOT) forums.slimdevices.com>:

>
> kdf Wrote:
> > I'll add a bit more in there hopefully for tomorrow's nightly build, to
> > track the scan trigger and we'll have to see the logs again.
>
> That sounds good. Unfortunately I'll be away for a week, but it looks
> like abunn is experiencing the same issue so may be able to provide
> feedback (assuming his logs look the same to start with). Otherwise
> I'll be back on to it next Monday.

np, I think I have some ideas. My library is very small on this machine so I was
not even noticing the startup scan, but I am seeing one now. However, it was
only one. Restarting the server has not triggered another scan yet.

> so I don't know what SS
> is doing on initialise (i.e. between start-up and me enabling the
> debug).

stop the service.

start->run->cmd

c:\program files\simserver\server\slim.exe --d_server --d_plugins --d_musicmagic
--d_remotestream

-kdf

hunta
2005-04-29, 16:05
Could be on to something here.

Every time I've started SS so far it's been from the launch icon, i.e. "C:\Program Files\SlimServer\SlimServer.exe". I just did your test, which involves starting slim.exe from the command line as a service. In the DOS window I could see the initialisation, then it whizzed through my library in a few seconds. SoftSqueeze was responsive the whole time. I can now navigate to localhost:9000 and everything's working 100% fine.

I'd now like to revert to starting the service using SlimServer.exe to check whether the problem's still there, but I can't work out how to stop it first..!

kdf
2005-04-29, 16:56
Quoting hunta <hunta.1o9w2n (AT) no-mx (DOT) forums.slimdevices.com>:

>
> Could be on to something here.
>
actually, I've noticed that the cacheid seems to be different every time
musicmagic is restarted, which will trigger the scan.

> I'd now like to revert to starting the service using SlimServer.exe to
> check whether the problem's still there, but I can't work out how to
> stop it first..!

ctrl-c :)

-kkdf

hunta
2005-04-29, 17:19
> actually, I've noticed that the cacheid seems to be different every time musicmagic is restarted, which will trigger the scan.

That fits with what I'm seeing here. I can close SS (ctrl-c!) and restart it from SlimServer.exe, and there's no problem at all. It's only when I close it down, close down and restart MMM that I see a problem. The difference I saw previously is that everything seems to run much faster and more smoothly from the DOS prompt than via SlimServer.exe - including scanning the library, but also pauses between plays etc.

So, even if there is a second underlying issue, one way and another it looks like it is better to run as a service. Does this have to run from a DOS shell in XP? I can't see it in my list of services (My Computer > Manage > Services).

Many thanks for all your help on this - most appreciated. Hopefully abunn will be able to back this up with what he sees too.

Andrew

kdf
2005-04-29, 17:29
On 29-Apr-05, at 5:19 PM, hunta wrote:
>
>
> So, even if there is a second underlying issue, one way and another it
> looks like it is better to run as a service. Does this have to run from
> a DOS shell in XP? I can't see it in my list of services (My Computer >
> Manage > Services)

slim.exe runs from command line, which gives you all of the feedback.
i really don't know much else about the particulars of the service
since I run slimserver.pl
on linux and windows. i can make that a service using firedaemon.
> .
>
> Many thanks for all your help on this - most appreciated. Hopefully
> abunn will be able to back this up with what he sees too.
>
happy to help, when there is something there to look at. hopefully
this actually leads to the solution :)

-k

abunn
2005-04-29, 17:51
Hey all:

I have the same symptom (snipprt of log file pasted below). MMM has finished
its analysis (0/0/0) and the scan starts about a minute after slim.exe is
run. My reload interval is set to 3600. I'm using SS 6.0.2 and MMM 1.1.5 One
thing I see here is that slim is reading 9660 songs in the MMM db.
Slimserver sees 9667 songs. The seven missing songs are too short for MMM to
analyze. What that says to me is that slim is making a read of the MMM db
successfully.

I'll leave slim running with MMM overnight and see if it finshes scanning
and is stable. It's hard to leave it running and inoperable (e.g, with the
CPU pegged) as the SB gets a lot of use.


Thanks kdf. You rock. -Andy

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~

Here are two starts where I started slim.exe with MMM already running.

Log file 1:

Use of uninitialized value in exists at /PerlApp/Slim/Buttons/Home.pm line
159.
Use of uninitialized value in exists at /PerlApp/Slim/Buttons/Home.pm line
159.
2005-04-29 20:30:21.7327 MusicMagic: Testing for API on localhost:10002
2005-04-29 20:30:21.9215 MusicMagic: Predixis MusicMagic Mixer. Version
1.1.5

2005-04-29 20:31:22.0244 MusicMagic: music library has changed!
2005-04-29 20:31:22.0245 MusicMagic: using musicmagic: Predixis MusicMagic
Mixer. Version 1.1.5

2005-04-29 20:31:22.0246 MusicMagic: start export
2005-04-29 20:31:22.0260 MusicMagic: export mode is: start
2005-04-29 20:31:22.1181 MusicMagic: Got 9660 song(s).
2005-04-29 20:31:22.1183 MusicMagic: export mode is: songs
2005-04-29 20:31:22.2124 MusicMagic: Exporting song 1:
F:\music\mp3s\1.8.7\01 - Worlds Apart.mp3
2005-04-29 20:31:22.2369 MusicMagic: export mode is: songs
....and so on.



Log file 2:

Use of uninitialized value in exists at /PerlApp/Slim/Buttons/Home.pm line
159.
Use of uninitialized value in exists at /PerlApp/Slim/Buttons/Home.pm line
159.
2005-04-29 20:38:36.6708 MusicMagic: Testing for API on localhost:10002
2005-04-29 20:38:36.7728 MusicMagic: Predixis MusicMagic Mixer. Version
1.1.5

2005-04-29 20:39:36.8550 MusicMagic: music library has changed!
2005-04-29 20:39:36.8552 MusicMagic: using musicmagic: Predixis MusicMagic
Mixer. Version 1.1.5

2005-04-29 20:39:36.8553 MusicMagic: start export
2005-04-29 20:39:36.8559 MusicMagic: export mode is: start
2005-04-29 20:39:36.9488 MusicMagic: Got 9660 song(s).
2005-04-29 20:39:36.9491 MusicMagic: export mode is: songs
2005-04-29 20:39:37.0432 MusicMagic: Exporting song 1:
F:\music\mp3s\1.8.7\01 - Worlds Apart.mp3
2005-04-29 20:39:37.0677 MusicMagic: export mode is: songs
....and so on.

kdf
2005-04-29, 18:38
On 29-Apr-05, at 5:51 PM, Andy Bunn wrote:
>
>
> Thanks kdf. You rock. -Andy
>

thanks, but I haven't actually solved anything yet. i've added a
couple new debug lines just now. if you use svn, please update. if
you use the windows exe, please download and install the 6.1 nightly
build for Apr 30 and see what you find. My current theory is that
musicmagic is creating a new cacheid on each startup, so i'm looking
for output from more than one run so that I can confirm this.

i have also set my reload interval for musicmagic to 0 and this avoids
the scan even on startup, so try that as well since that may at least
lessen the problem in the short term.

-kdf

abunn
2005-04-29, 19:26
Alright. I let slim rescan (it took less than an hour). I could then kill
slim and restart it without any problem. When I restarted MMM and ran slim
it scanned again:

Use of uninitialized value in exists at /PerlApp/Slim/Buttons/Home.pm line
159.
Use of uninitialized value in exists at /PerlApp/Slim/Buttons/Home.pm line
159.
2005-04-29 21:49:06.3227 MusicMagic: Testing for API on localhost:10002
2005-04-29 21:49:06.3414 Executing command 00:04:20:05:71:8c: playlist
(add) (F:\music\slimplaylists\__00_04_20_05_71_8c.m3u) () () () () ()
2005-04-29 21:49:06.5265 Returning array: playlist (add)
(F:\music\slimplaylists\__00_04_20_05_71_8c.m3u) () () () () ()
2005-04-29 21:49:06.5310 MusicMagic: Predixis MusicMagic Mixer. Version
1.1.5

2005-04-29 21:49:34.2200 Executing command no client: rescan () () () () ()
() ()
2005-04-29 21:49:34.5890 MusicMagic: using musicmagic: Predixis MusicMagic
Mixer. Version 1.1.5

2005-04-29 21:49:34.5893 MusicMagic: start export
2005-04-29 21:49:34.5894 Returning array: rescan () () () () () () ()
2005-04-29 21:49:34.6716 MusicMagic: export mode is: start
2005-04-29 21:49:34.7715 MusicMagic: Got 9660 song(s).
2005-04-29 21:49:34.8053 MusicMagic: export mode is: songs
2005-04-29 21:49:34.9229 MusicMagic: Exporting song 1:
F:\music\mp3s\1.8.7\01 - Worlds Apart.mp3
2005-04-29 21:49:34.9519 MusicMagic: export mode is: songs
....and so on.


So, when MMM is initialized, slim feels compelled to scan (orwhatever the
right word is).

-Andy

kdf
2005-04-29, 20:19
On 29-Apr-05, at 7:26 PM, Andy Bunn wrote:

>
>
>
> So, when MMM is initialized, slim feels compelled to scan (orwhatever
> the
> right word is).
>
that is the current theory. try tomorrow's build for a bit more debug
around that area. it wont change anything, but it will at least give
recorded proof of the variables that determine when to scan. that way
I can be sure I'm not just making it up :)

wendell is also working from the Predixis side to give is a bit more to
work with to determine when is and is not a good time to scan.

-kdf

hunta
2005-04-30, 00:30
i have also set my reload interval for musicmagic to 0 and this avoids the scan even on startup, so try that as well since that may at least lessen the problem in the short term.

That doesn't seem to be the case when I'm starting SS as SlimServer.exe. I've had the reload interval set to 0 for some time now (since post #45).


that is the current theory. try tomorrow's build for a bit more debug around that area. it wont change anything, but it will at least give recorded proof of the variables that determine when to scan. that way I can be sure I'm not just making it up :)
Again, sorry that I won't be around for a week now. I'll get straight back on to it next Monday.


wendell is also working from the Predixis side to give is a bit more to work with to determine when is and is not a good time to scan.
Thanks to wendell to for your help.

Good luck!

Andrew

kdf
2005-04-30, 00:38
Quoting hunta <hunta.1oaj80 (AT) no-mx (DOT) forums.slimdevices.com>:

>
> kdf Wrote:
> > i have also set my reload interval for musicmagic to 0 and this avoids
> > the scan even on startup, so try that as well since that may at least
> > lessen the problem in the short term.
>
> That doesn't seem to be the case when I'm starting SS as
> SlimServer.exe. I've had the reload interval set to 0 for some time now
> (since post #45).

make sure you have the latest of the nightly builds for 6.1 (NOT 6.0.3), April
28 or newer.

-k

abunn
2005-04-30, 18:06
kdf wrote:
> make sure you have the latest of the nightly builds for 6.1 (NOT 6.0.3),
April 28 or newer.

I installed the latest nightly (4/30) and see the new info being dumped to
the log file. See below. It's complaining abput a "uninitialized value in
concatenation" but seems to run. I can use MMM, make mixes, etc. and slim is
not rescanning. It does see a new cache id and the library has not changed.

Curious.

-Andy

~~~~~~~~~~~~~~~~~~~~~~~~~

Subroutine new redefined at C:/Program
Files/SlimServer/server/Plugins/RadioIO/Plugin.pm line 239.
2005-04-30 20:56:32.7061 MusicMagic: Testing for API on localhost:10002
2005-04-30 20:56:32.7288 MusicMagic: Predixis MusicMagic Mixer. Version
1.1.5

2005-04-30 20:56:33.1074 Executing command 00:04:20:05:71:8c: playlist
(add) (F:\music\slimplaylists\__00_04_20_05_71_8c.m3u) () () () () ()
2005-04-30 20:56:33.2862 Returning array: playlist (add)
(F:\music\slimplaylists\__00_04_20_05_71_8c.m3u) () () () () ()
2005-04-30 20:57:32.7767 MusicMagic: read cacheid of 1114908768

2005-04-30 20:57:32.7770 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 20:57:32.7772 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 20:57:32.7773 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 20:57:44.7698 Musicmagic request:
http://localhost:10002/api/mix?artist%3DThe%20Allman%20Brothers%20Band&varie
ty=0&style=0&sizetype=tracks&size=12
....and so on.

kdf
2005-04-30, 18:18
Quoting Andy Bunn <abunn (AT) whrc (DOT) org>:

> Curious.

thanks. the uninitialized value warnings are just warnings and harmless. They
are coming up because the debug asks for the last scan time. since you havent'
scanned since startup, its indefined.

with 6.1 nightlies, a zero in the reload interval will prevent the scan, as the
log shows. The cacheid issue is something I have sent off to Predixis as a
query. It seems that shutting down with unfinished business can cause the next
start to have a new cacheid. I dont yet know what that unfinished business
might be. at least now, with 6.1 you can leave your reload interval set to 0
and only reload when you want to manually or via the rescan plugin (auto rescan
every 24 hours and a set time)

-kdf

abunn
2005-04-30, 19:14
Excellent. Thanks for the translation. And I forgot to post this, but you
probably already know, the recan attempts are continuous. I think the MMM
plugin reports a change every two seconds or so.

-A


~~~~~~~~~~~~~~~~~~~~~~~~~~

2005-04-30 21:45:04.7145 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:45:04.7146 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:45:04.7147 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:46:04.8100 MusicMagic: read cacheid of 1114908768

2005-04-30 21:46:04.8103 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:46:04.8104 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:46:04.8105 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:47:04.9017 MusicMagic: read cacheid of 1114908768

2005-04-30 21:47:04.9020 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:47:04.9021 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:47:04.9022 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:48:05.0111 MusicMagic: read cacheid of 1114908768

2005-04-30 21:48:05.0114 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:48:05.0115 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:48:05.0115 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:49:05.0892 MusicMagic: read cacheid of 1114908768

2005-04-30 21:49:05.0895 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:49:05.0896 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:49:05.0897 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:50:05.1673 MusicMagic: read cacheid of 1114908768

2005-04-30 21:50:05.1676 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:50:05.1677 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:50:05.1678 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:51:05.2610 MusicMagic: read cacheid of 1114908768

2005-04-30 21:51:05.2613 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:51:05.2614 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:51:05.2615 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:52:05.3548 MusicMagic: read cacheid of 1114908768

2005-04-30 21:52:05.3551 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:52:05.3552 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:52:05.3553 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:53:05.4486 MusicMagic: read cacheid of 1114908768

2005-04-30 21:53:05.4489 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:53:05.4490 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -
2005-04-30 21:53:05.4491 MusicMagic: Scan Interval set to 0, rescanning
disabled
2005-04-30 21:54:05.5111 MusicMagic: read cacheid of 1114908768

2005-04-30 21:54:05.5114 MusicMagic: music library has changed!
Use of uninitialized value in concatenation (.) or string at C:/Program
Files/SlimServer/server/Plugins/MusicMagic/Plugin.pm line 239.
2005-04-30 21:54:05.5115 Details:
Cacheid - 1114908768

LastCacheid - 1114830847
Reload Interval - 0
Last Scan -

kdf
2005-04-30, 19:47
On 30-Apr-05, at 7:14 PM, Andy Bunn wrote:

> Excellent. Thanks for the translation. And I forgot to post this, but
> you
> probably already know, the recan attempts are continuous. I think the
> MMM
> plugin reports a change every two seconds or so.
>
> -A
>
the checker runs every 60 seconds. You won't notice it if you turn the
d_musicmagic debug flag off :)
no rescan is attempted, and only one connect to musicmagic API is done
during this time.

-k