PDA

View Full Version : sb2 slimserver version



Kirk Ferguson
2005-03-11, 13:59
Hello. I'm wondering whether the new squeezboxes will work with older versions
of Slimserver, say 5.4.0 or 5.4.1. I'm planning to order one if so.

I've had terrible luck so far with version 6, and wouldn't want to be locked
into using it until I can get it running.

After trying various nightlies recently, I've finally gotten the new beta to
scan my library, but once the scan completes (3 hours) the server crashes
(without log entries) as soon as I try to browse using the web interface. To
date I've never actually gotten version 6 to play a single note.

Fortunately, 5.4.x runs fine on the same system (dual P3, SCSI, 1GB, RedHat EL),
although 5.4.1 produces skips and occasionally cuts off songs early.

Seems that for my environment Slimserver is regressing over time, so I hope I
can use the more stable version with the new hardware.

Kirk

kdf
2005-03-11, 14:13
Quoting Kirk Ferguson <kferguson (AT) students (DOT) cncc.edu>:

> Hello. I'm wondering whether the new squeezboxes will work with older
> versions
> of Slimserver, say 5.4.0 or 5.4.1. I'm planning to order one if so.

6.0 will be required. The squeezebox 2 is totally new hardware, so it needs
special new code to talk to it.

> I've had terrible luck so far with version 6, and wouldn't want to be locked
> into using it until I can get it running.
>
> After trying various nightlies recently, I've finally gotten the new beta to
> scan my library, but once the scan completes (3 hours) the server crashes
> (without log entries) as soon as I try to browse using the web interface. To
> date I've never actually gotten version 6 to play a single note.

3 hours seems long. How many songs?

> Fortunately, 5.4.x runs fine on the same system (dual P3, SCSI, 1GB, RedHat
> EL),

try running /usr/local/slimserver.pl --d_http --d_info --d_source

you'll see the web activity, song infomation handling and playback processing in
the terminal. When/if the crash occurs, you can report the tail end of that
log.

-kdf

Kirk Ferguson
2005-03-11, 14:25
Quoting kdf <slim-mail (AT) deane-freeman (DOT) com>:

> Quoting Kirk Ferguson <kferguson (AT) students (DOT) cncc.edu>:
>
> > Hello. I'm wondering whether the new squeezboxes will work with older
> > versions
> > of Slimserver, say 5.4.0 or 5.4.1. I'm planning to order one if so.
>
> 6.0 will be required. The squeezebox 2 is totally new hardware, so it needs
> special new code to talk to it.
>

Bummer...

> > I've had terrible luck so far with version 6, and wouldn't want to be
> locked
> > into using it until I can get it running.
> >
> > After trying various nightlies recently, I've finally gotten the new beta
> to
> > scan my library, but once the scan completes (3 hours) the server crashes
> > (without log entries) as soon as I try to browse using the web interface.
> To
> > date I've never actually gotten version 6 to play a single note.
>
> 3 hours seems long. How many songs?

Around 25,000.
>
> > Fortunately, 5.4.x runs fine on the same system (dual P3, SCSI, 1GB,
> RedHat
> > EL),
>
> try running /usr/local/slimserver.pl --d_http --d_info --d_source
>
> you'll see the web activity, song infomation handling and playback processing
> in
> the terminal. When/if the crash occurs, you can report the tail end of that
> log.
>
I'll try that once I install version 6 again. Takes over three hours each time
I install, scan and try it out. Right now I think I'd just like to listen to
music for a while now that I've backed up to 5.4 again.

Thanks for the help. I'll post the output once I get it.

> -kdf
>

kdf
2005-03-11, 14:32
Quoting Kirk Ferguson <kferguson (AT) students (DOT) cncc.edu>:

> > 3 hours seems long. How many songs?
>
> Around 25,000.
> >

it would be very useful to have a snippet of the d_info log from during the
scan. I have about 10,000 in my collection and it takes a mere 15-18 minutes
from a blank slate. 35-40 minutes if I have itunes and musicmagic enabled at
the same time.

When you do get a chance to try 6.0 again, be sure to grab the latest nightly
build; improvements are coming in daily. Also, if you do complete a scan, keep
the slimserversql.db file around. You can save yourself a rescan if you do.
Plus, Dan may want to have a look at it if you are still having problems.

cheers,
kdf

JJ
2005-03-11, 15:04
----- Original Message -----
From: "kdf" <slim-mail (AT) deane-freeman (DOT) com>
To: "Slim Devices Discussion" <discuss (AT) lists (DOT) slimdevices.com>
Sent: Friday, March 11, 2005 2:13 PM
Subject: [slim] sb2 slimserver version


> Quoting Kirk Ferguson <kferguson (AT) students (DOT) cncc.edu>:
>
>> Hello. I'm wondering whether the new squeezboxes will work with older
>> versions
>> of Slimserver, say 5.4.0 or 5.4.1. I'm planning to order one if so.
>
>6.0 will be required. The squeezebox 2 is totally new hardware, so it
>needs
>special new code to talk to it.


Slim Devices is going to be shipping a new product in less than three
weeks and the software for it is still be in beta? Of course we can all
expect 6.0 to be "released" by then, but given all the problems I'm
reading about there seems very little chance they'll be ironed out in 20
days.

This should be interesting.

Roy Owen
2005-03-11, 15:11
From what I've seen so far I expect that SD will be able to meet the
deadline with functional
server software. I've gotten very good responses to my problems and
remember this is a b1 not a b0 release. The folks at SD are dedicated
and take pride in their work, so don't sell them short. Also the user
base is more "tech savvy" than the norm (read other geeks/developers)
so the feedback they're getting is *VERY* good, also loyal.

Just my $0.02

Roy

On Fri, 11 Mar 2005 15:04:24 -0700, JJ <jj (AT) zolx (DOT) com> wrote:
> ----- Original Message -----
> From: "kdf" <slim-mail (AT) deane-freeman (DOT) com>
> To: "Slim Devices Discussion" <discuss (AT) lists (DOT) slimdevices.com>
> Sent: Friday, March 11, 2005 2:13 PM
> Subject: [slim] sb2 slimserver version
>
> > Quoting Kirk Ferguson <kferguson (AT) students (DOT) cncc.edu>:
> >
> >> Hello. I'm wondering whether the new squeezboxes will work with older
> >> versions
> >> of Slimserver, say 5.4.0 or 5.4.1. I'm planning to order one if so.
> >
> >6.0 will be required. The squeezebox 2 is totally new hardware, so it
> >needs
> >special new code to talk to it.
>
> Slim Devices is going to be shipping a new product in less than three
> weeks and the software for it is still be in beta? Of course we can all
> expect 6.0 to be "released" by then, but given all the problems I'm
> reading about there seems very little chance they'll be ironed out in 20
> days.
>
> This should be interesting.
>
>
>

Steve Bernard, Jr
2005-03-11, 15:19
On Fri, 11 Mar 2005 13:32:04 -0800, kdf wrote:
> Quoting Kirk Ferguson <kferguson (AT) students (DOT) cncc.edu>:
>
> > > 3 hours seems long. How many songs?
> >
> > Around 25,000.
> > >
>
> it would be very useful to have a snippet of the d_info log from during the
> scan. I have about 10,000 in my collection and it takes a mere 15-18 minutes
> from a blank slate. 35-40 minutes if I have itunes and musicmagic enabled at
> the same time.

FWIW, my library of 10,400 or so files takes somewhere around four
hours to scan from a blank slate on 6.0, using the latest version.

My library is almost completely Ogg Vorbis files, but I don't know if
that makes a difference. I'll check that debug log as well next time
I rescan.

-Steve

Dan Sully
2005-03-11, 15:21
* Steve Bernard, Jr shaped the electrons to say...

>FWIW, my library of 10,400 or so files takes somewhere around four
>hours to scan from a blank slate on 6.0, using the latest version.
>
>My library is almost completely Ogg Vorbis files, but I don't know if
>that makes a difference. I'll check that debug log as well next time I rescan.

Steve - what OS are you using? If you can compile Ogg::Vorbis::Header from
CPAN - which a compiled C version, instead of the pure perl version which we
ship with SlimServer. That should be much faster for you.

-D
--
This knob controls the thing that changes when you turn it. - noah

Gregory Hamilton
2005-03-11, 16:17
Steve,

I was suprised to see your benchmark. I would like to provide one for
comparison:

Specs:
2GHZ Laptop
Windows XP SP1
17259 MP3's on a USB 1.1 harddrive.
Moodlogic enabled

Total time for new database build was 50 minutes.



On Fri, 11 Mar 2005 14:21:47 -0800, Dan Sully <dan (AT) slimdevices (DOT) com> wrote:
> * Steve Bernard, Jr shaped the electrons to say...
>
> >FWIW, my library of 10,400 or so files takes somewhere around four
> >hours to scan from a blank slate on 6.0, using the latest version.
> >
> >My library is almost completely Ogg Vorbis files, but I don't know if
> >that makes a difference. I'll check that debug log as well next time I rescan.
>
> Steve - what OS are you using? If you can compile Ogg::Vorbis::Header from
> CPAN - which a compiled C version, instead of the pure perl version which we
> ship with SlimServer. That should be much faster for you.
>
> -D
> --
> This knob controls the thing that changes when you turn it. - noah
>

Mike Kozlowski
2005-03-11, 17:45
On Fri, 11 Mar 2005, JJ wrote:

> Slim Devices is going to be shipping a new product in less than three weeks
> and the software for it is still be in beta?

That's pretty much how it worked for the Squeezebox G, too. Which is why
I held off ordering the display upgrade for most of a month after it was
released, and why I'll hold off on a Squeezebox 2 upgrade for a while
after its release. Let others be gamma testers, I say.

--
Mike Kozlowski
http://www.klio.org/mlk/

Steve Bernard, Jr
2005-03-11, 18:37
On Fri, 11 Mar 2005 14:21:47 -0800, Dan Sully wrote:
> * Steve Bernard, Jr shaped the electrons to say...
>
> >FWIW, my library of 10,400 or so files takes somewhere around four
> >hours to scan from a blank slate on 6.0, using the latest version.
> >
> >My library is almost completely Ogg Vorbis files, but I don't know if
> >that makes a difference. I'll check that debug log as well next time I rescan.
>
> Steve - what OS are you using? If you can compile Ogg::Vorbis::Header from
> CPAN - which a compiled C version, instead of the pure perl version which we
> ship with SlimServer. That should be much faster for you.
>

I'm running Ubuntu Linux as of fairly recently. I compiled by own
Ogg::Vorbis::Header and, holy cow, a wipe and rescan took 38 minutes!
That's quite an improvement.

I even wiped and rescanned again just to watch the DB fill up. Such
is my sad, sad life.

Thanks,

-Steve

P.S. I'm still occassionally getting a No Album entry in my DB,
usually if I kill and restart the server, but I can't reproduce it
reliably enough to give you any more info. It's nothing like it was
before on 6.0, though.

Dan Sully
2005-03-11, 19:23
* Steve Bernard, Jr shaped the electrons to say...

>I'm running Ubuntu Linux as of fairly recently. I compiled by own
>Ogg::Vorbis::Header and, holy cow, a wipe and rescan took 38 minutes!
>That's quite an improvement.

Great! Eventually I'd like to include Ogg::Vorbis::Header, but because it
relies on the vorbis libraries, we need to get all that working with dynamic
libraries, etc.

>I even wiped and rescanned again just to watch the DB fill up. Such
>is my sad, sad life.

Heh

>P.S. I'm still occassionally getting a No Album entry in my DB,
>usually if I kill and restart the server, but I can't reproduce it
>reliably enough to give you any more info. It's nothing like it was
>before on 6.0, though.

Does the No Album entry have tracks listed under it? That entry is created if
SlimServer finds a file which doesn't have an ALBUM tag.

-D
--
<iNoah> you know, most free operating systems come preinstalled with their own high horse.

Steve Bernard, Jr
2005-03-12, 10:23
On Fri, 11 Mar 2005 18:23:35 -0800, Dan Sully wrote:

> Great! Eventually I'd like to include Ogg::Vorbis::Header, but because it
> relies on the vorbis libraries, we need to get all that working with dynamic
> libraries, etc.
>

Just compiling this for myself was no easy feat. First it bombed out
looking for MakeMaker.pm, so I installed one, but it actually wanted a
*different* MakeMaker.pm. Once that was solved, it failed looking for
libogg and libvorbis. I knew I had those installed, but it turned out
I also needed libogg-dev and libvorbis-dev. Seems like a chore to
make sure this sort of thing Just Works for whoever downloads
SlimServer.

> >P.S. I'm still occassionally getting a No Album entry in my DB,
> >usually if I kill and restart the server, but I can't reproduce it
> >reliably enough to give you any more info. It's nothing like it was
> >before on 6.0, though.
>
> Does the No Album entry have tracks listed under it? That entry is created if
> SlimServer finds a file which doesn't have an ALBUM tag.

I tested this a bit more this morning and I've got what seems to be a
repeatable way to get a No Album and No Artist entry in my DB.

1. Start the server with a fresh rescan. Album count is 896 at this
point for me, Artists is 890.

2. Go to the Internet Radio link under Help, and enter an MP3 url into
the field. For tests, I've been using the MP3 samples on the Darla
Records home page, such as http://www.darla.com/mp3/mp39026.mp3. In
the Now Playing window the track title is the URL. The artist is No
Artist and the album is No Album.

3. Let the MP3 play for ten seconds or so then go back to the main
page. Album count is now 897 with 891 artists. A No Artist and a No
Album entry appear in the DB. Browse Artists -> No Artist shows "1
album with 1 song by 1 artist". Browsing to the album "No Album" has
no tracks in the list, although the web interface says there should be
1 song. While the album and artist counts have gone up, the number of
songs remains steady at 10434.

4. If I go back to Internet Radio and play more URLs, the number of
songs that are supposedly under No Artist -> No Album increases
accordingly. No tracks actually appear in the listing and the
'official' number of songs in the DB doesn't increase.

So, it looks like SlimServer is adding track info for songs not
actually in my library to the DB. A rescan doesn't seem to make the
server realize they're not really there.

Let me know if you need this info in Bugzilla in an new or existing bug.

-Steve

Dan Sully
2005-03-12, 10:58
* Steve Bernard, Jr shaped the electrons to say...

>Just compiling this for myself was no easy feat. First it bombed out
>looking for MakeMaker.pm, so I installed one, but it actually wanted a
>*different* MakeMaker.pm. Once that was solved, it failed looking for
>libogg and libvorbis. I knew I had those installed, but it turned out
>I also needed libogg-dev and libvorbis-dev. Seems like a chore to
>make sure this sort of thing Just Works for whoever downloads
>SlimServer.

Yes - it's kind of a pain. Which is why we'd like to include it. The pure
perl module is rather unoptimized, as you found out. But it works everywhere.

>4. If I go back to Internet Radio and play more URLs, the number of
>songs that are supposedly under No Artist -> No Album increases
>accordingly. No tracks actually appear in the listing and the
>'official' number of songs in the DB doesn't increase.
>
>So, it looks like SlimServer is adding track info for songs not
>actually in my library to the DB. A rescan doesn't seem to make the
>server realize they're not really there.
>
>Let me know if you need this info in Bugzilla in an new or existing bug.

What version of the server is this with?

Thanks.

-D
--
They're techno trousers, ex-NASA, fantastic for walkies!

Steve Bernard, Jr
2005-03-12, 11:12
> What version of the server is this with?
>
> Thanks.
>
> -D

This was a Subversion build I put together this morning. The 2447
checkin just happened before I got back to the computer, so I'm
guessing 2446 or thereabouts.

-Steve

Avi Schwartz
2005-03-12, 17:45
Dan Sully wrote:

> * Steve Bernard, Jr shaped the electrons to say...
>
>> FWIW, my library of 10,400 or so files takes somewhere around four
>> hours to scan from a blank slate on 6.0, using the latest version.
>>
>> My library is almost completely Ogg Vorbis files, but I don't know if
>> that makes a difference. I'll check that debug log as well next time
>> I rescan.
>
>
> Steve - what OS are you using? If you can compile Ogg::Vorbis::Header
> from
> CPAN - which a compiled C version, instead of the pure perl version
> which we
> ship with SlimServer. That should be much faster for you.
>
> -D

Are there any instructions/steps on how to compile Ogg::Vorbis::Header
on a Linux box? and would it also speed up scanning FLAC files (which I
believe also use the Ogg tags)?

Thanks,

Avi

--
Avi Schwartz
http://public.xdi.org/=avi.schwartz

Dan Sully
2005-03-12, 17:46
* Avi Schwartz shaped the electrons to say...

>>Steve - what OS are you using? If you can compile Ogg::Vorbis::Header
>>from
>>CPAN - which a compiled C version, instead of the pure perl version
>>which we
>>ship with SlimServer. That should be much faster for you.
>>
>Are there any instructions/steps on how to compile Ogg::Vorbis::Header
>on a Linux box? and would it also speed up scanning FLAC files (which I
>believe also use the Ogg tags)?

What Linux distribution are you on? Some have it available as a package.

You'll need libvorbis and libogg.

FLAC, although it uses vorbis tags, has a different header format. I wrote
Audio::FLAC::Header - which has both a compiled and pure perl part. It will
fall back to the perl only if the XS module isn't available. You can compile
it for more speed. libFLAC is needed.

-D
--
<iNoah> my pdp goes to 11.

Avi Schwartz
2005-03-12, 17:56
Dan Sully wrote:

> * Avi Schwartz shaped the electrons to say...
>
>>> Steve - what OS are you using? If you can compile
>>> Ogg::Vorbis::Header from
>>> CPAN - which a compiled C version, instead of the pure perl version
>>> which we
>>> ship with SlimServer. That should be much faster for you.
>>>
>> Are there any instructions/steps on how to compile
>> Ogg::Vorbis::Header on a Linux box? and would it also speed up
>> scanning FLAC files (which I believe also use the Ogg tags)?
>
>
> What Linux distribution are you on? Some have it available as a package.
>
Mandrake 10.1.

Since I am using only FLAC with the occasional MP3, I guess I don't need
to compile Ogg::Vorbis::Header.

> FLAC, although it uses vorbis tags, has a different header format. I
> wrote
> Audio::FLAC::Header - which has both a compiled and pure perl part. It
> will
> fall back to the perl only if the XS module isn't available. You can
> compile
> it for more speed. libFLAC is needed.

I believe I have XS installed since I have the
/usr/lib/perl5/5.8.5/i386-linux-thread-multi/XS and auto/XS directories.

I also have libFLAC installed.

So then, how do I compile it? I am afraid Perl is a mystery for me at
this point.

Avi

--
Avi Schwartz
http://public.xdi.org/=avi.schwartz

Dan Sully
2005-03-13, 00:17
* Avi Schwartz shaped the electrons to say...

>I also have libFLAC installed.
>
>So then, how do I compile it? I am afraid Perl is a mystery for me at this point.

Download http://www.cpan.org/authors/id/D/DA/DANIEL/Audio-FLAC-Header-1.4.tar.gz

Untar it, and then run 'perl Makefile.PL' in that directory.

make && make test && make install

-D
--
Do not panic, do not panic! We are trained professionals!
Now, stay calm. We are going around the leaf.