PDA

View Full Version : Server Upgrade Woes - Should have known better.



Michael Bowyer
2004-09-05, 23:50
Well, I can sympathise, I'm having the same issues myself. Having just
spent almost $400 on a new SB and a graphic upgrade, all I have is an
expensive exercise in frustration. Frequent server drop-outs, frequent
music drop-outs, and non-responding remote control buttons make the
system virtually unusable. Every couple of days there's a new beta
server version and firmware upgrade that fixes some issues, but usually
creates others. I'm sure Version 5 of slimserver has never been
stable.

Slim, please make it do what it says on the box - for non-geeks. The
hardware is far too expensive to be this problematic.

Depressed,
Mike.

Daniel Cohen
2004-09-06, 00:24
On 6/9/04 at 8:50 am +0200, Michael Bowyer wrote
>Well, I can sympathise, I'm having the same issues myself. Having
>just spent almost $400 on a new SB and a graphic upgrade, all I have
>is an expensive exercise in frustration. Frequent server drop-outs,
>frequent music drop-outs, and non-responding remote control buttons
>make the system virtually unusable. Every couple of days there's a
>new beta server version and firmware upgrade that fixes some issues,
>but usually creates others. I'm sure Version 5 of slimserver has
>never been stable.
>
>Slim, please make it do what it says on the box - for non-geeks. The
>hardware is far too expensive to be this problematic.

Question is what version of software and firmware are you using?
Ifind this list valuable but have decided *only* to use the public
releases, not the nightly versions. And I've had no significant
problems related to the product. I have some problems which are
clearly down to my wifi network, and some where the manual is not
clear as to how the Squeezebox works.

If you are a non-geek, stick to the public releases. The nightlies by
their nature have not been thoroughly tested, and part of the point
of making them available is to allow a wider base of systems to test
on.

But, agreed, it does seem that some of the bugs could be caught by
looking at the program before release.

--
Daniel Cohen

Robin Bowes
2004-09-06, 04:31
On Mon, September 6, 2004 7:50, Michael Bowyer said:
> Well, I can sympathise, I'm having the same issues myself. Having just
> spent almost $400 on a new SB and a graphic upgrade, all I have is an expensive exercise
> in frustration. Frequent server drop-outs, frequent music drop-outs, and non-responding
> remote control buttons make the system virtually unusable. Every couple of days there's
> a new beta server version and firmware upgrade that fixes some issues, but usually
> creates others. I'm sure Version 5 of slimserver has never been stable.
>
> Slim, please make it do what it says on the box - for non-geeks. The
> hardware is far too expensive to be this problematic.

Mike,

I can sympathise, to a degree, and I believe that there should be a "stable" version of
slimserver as well as a development branch.

The problem is that currently there doesn't seem to be a process to fix bugs in the
production code without jumping to a development version which often has many more
unresolved issues.

The stable version should be just that: stable; this would be the software that most
end-users run. There should be incremental releases to fix bugs and some new features
(if extensively tested).

The development branch should be where all new features are first implemented and
tested. It will most likely be less stable than the "stable" branch, but there should
still be formal releases that are fairly stable and usable "out-of-the-box".

For the real geeks (or masochists) there is always CVS (or the latest nightly builds)
but anyone using these releases should not moan when things break!

Just my thoughts...

R.
--
http://robinbowes.com

Robin Bowes
2004-09-06, 08:01
On Mon, September 6, 2004 15:25, Phillip Kerman said:
>> Question is what version of software and firmware are you using?
>> Ifind this list valuable but have decided *only* to use the public
>> releases, not the nightly versions. And I've had no significant problems related to the
>> product. I have some problems which are clearly down to my wifi network, and some
>> where the manual is not clear as to how the Squeezebox works.
>
> But the graphic display only works with the beta of 5.3

Good point.

So, slim devices are currently selling a product that only works with beta software -
not good.

To re-iterate my previous comments, there should be two distinct versions of slimserver:

1. stable or release version, wich is updated to fix bugs and introduce new features
after they have been thoroughly tested, e.g. support for the graphic display.

2. development or beta version, which is where most bugs are fixed and tested and where
new features are first introduced and tested.

Plus of course:

3. CVS (or nightly builds), for those who like to be bleeding edge!

This is how a lot of (most?) software is distributed and should keep most people happy.

Non-techy users can use the latest stable version - less features but stable
Geeks and hackers can use the latest development version - latest features, but probably
more bugs.
And masochists can update from CVS or use the latest nightly build!

How about it?

R.
--
http://robinbowes.com

Jack Coates
2004-09-06, 08:01
>> Question is what version of software and firmware are you using?
>> Ifind this list valuable but have decided *only* to use the public
>> releases, not the nightly versions. And I've had no significant
>> problems related to the product. I have some problems which are
>> clearly down to my wifi network, and some where the manual is not
>> clear as to how the Squeezebox works.
>>
>
>
> But the graphic display only works with the beta of 5.3
>
> Thanks,
> Phillip

that's not true -- you only get the purty fonts with the 5.3 beta, but the
display is still entirely functional with the released software version.

--
Jack At Monkeynoodle.Org: It's A Scientific Venture...
"Believe what you're told; there'd be chaos if everyone thought for
themselves." -- Top Dog hotdog stand, Berkeley, CA

Steve Baumgarten
2004-09-06, 08:30
> But the graphic display only works with the beta of 5.3

I think this is a legitimate complaint. The SBG went out the door too
soon. I was confused when I saw the latest release as 5.2.1 on the web
site even after the SBG upgrades had started to ship (I had just gotten
mine and new I'd need a new server version). In fact 5.2.1 is still listed
as the release version on the web site even though if you have an SBG that
version won't work with it.

I'm on this list so I know about the nightlies; on the other hand, the
nightlies are sometimes good, sometimes not so good (which goes with the
territory), so I wasn't thrilled to discover I'd be playing nightly
roulette in order to continue to be able to use my SB with the G upgrade.

Things are now in a pretty decent state with a nightly from the last few
days and firmware 35; but there was a fair amount of time between the
arrival of my G upgrade and the arrival of a stable and relatively
non-buggy server/firmware combo (which in any event still isn't a
"release" version).

If people are now buying SBGs in retail boxes, they're going to have not
such great experiences with them.

That's -- well, that's not so great.

For myself: I've always treated the SB as a techie thing. I'm a techie; I
enjoy following the SB development process. I can deal with helping debug
things. On the other hand I've just recently shown my SBG off to a couple
of friends, both decidedly non-techie. What happens when they buy one?

SBB

Jack Coates
2004-09-06, 08:40
....
>
> So, slim devices are currently selling a product that only works with beta
> software -
> not good.
>

Perhaps you're new to the computer business, or suffering temporary amnesia?

> To re-iterate my previous comments, there should be two distinct versions
> of slimserver:
>
> 1. stable or release version, wich is updated to fix bugs and introduce
> new features
> after they have been thoroughly tested, e.g. support for the graphic
> display.

Who's going to test these? Certainly not the people complaining about
bugs. QA costs a lot more money than it brings in. Ditto for tech support.
Yeah, that sucks, but that's the way things are in the computer business.
This is the point where you'll argue that Slim Devices is entering the
consumer electronics business, a bastion of reliability, to which I must
laugh.

>
> 2. development or beta version, which is where most bugs are fixed and
> tested and where
> new features are first introduced and tested.
>
> Plus of course:
>
> 3. CVS (or nightly builds), for those who like to be bleeding edge!
>
> This is how a lot of (most?) software is distributed and should keep most
> people happy.
>

examples, please. The only one I can think of is Debian.

> Non-techy users can use the latest stable version - less features but
> stable

like say, the released version, 5.2.1?

> Geeks and hackers can use the latest development version - latest
> features, but probably
> more bugs.

Hm, you must be referring to the beta version, 5.3.0b1.

> And masochists can update from CVS or use the latest nightly build!
>

downloads/nightly/latest!

> How about it?

If you're looking for completely separate trees, go ahead and fork, the
code is open. If you need some help setting up a CVS server, let me know
off-list. You might want to investigate what's already in progress though,
the database-backed branch is not the one you want.

--
Jack At Monkeynoodle.Org: It's A Scientific Venture...
"Believe what you're told; there'd be chaos if everyone thought for
themselves." -- Top Dog hotdog stand, Berkeley, CA

Jack Coates
2004-09-06, 08:44
>> But the graphic display only works with the beta of 5.3
>
> I think this is a legitimate complaint. The SBG went out the door too
> soon. I was confused when I saw the latest release as 5.2.1 on the web
> site even after the SBG upgrades had started to ship (I had just gotten
> mine and new I'd need a new server version). In fact 5.2.1 is still listed
> as the release version on the web site even though if you have an SBG that
> version won't work with it.
>

This is not true. You'll get more featureful display by using the beta,
but you will not be left with a doorstop if you use the stable version.

> I'm on this list so I know about the nightlies; on the other hand, the
> nightlies are sometimes good, sometimes not so good (which goes with the
> territory), so I wasn't thrilled to discover I'd be playing nightly
> roulette in order to continue to be able to use my SB with the G upgrade.
>

so don't. Use the stable version and wait for other people to test the new
features for you.

> Things are now in a pretty decent state with a nightly from the last few
> days and firmware 35; but there was a fair amount of time between the
> arrival of my G upgrade and the arrival of a stable and relatively
> non-buggy server/firmware combo (which in any event still isn't a
> "release" version).
>
> If people are now buying SBGs in retail boxes, they're going to have not
> such great experiences with them.
>
> That's -- well, that's not so great.
>
> For myself: I've always treated the SB as a techie thing. I'm a techie; I
> enjoy following the SB development process. I can deal with helping debug
> things. On the other hand I've just recently shown my SBG off to a couple
> of friends, both decidedly non-techie. What happens when they buy one?
>
Venkman: This city is headed for a disaster of biblical proportions.
Mayor: What do you mean, biblical?
Ray: What he means is Old Testament, Mr. Mayor... real
Wrath-of-God-type stuff. Fire and brimstone coming down from the
skies.
Venkman: Rivers and seas boiling!
Egon: 40 years of darkness, earthquakes, volcanos.
Winston:The dead rising from the grave!
Venkman: Human sacrifice, dogs and cats, living together... mass
hysteria!

--
Jack At Monkeynoodle.Org: It's A Scientific Venture...
"Believe what you're told; there'd be chaos if everyone thought for
themselves." -- Top Dog hotdog stand, Berkeley, CA

Steve Baumgarten
2004-09-06, 08:44
> that's not true -- you only get the purty fonts with the 5.3 beta, but
> the display is still entirely functional with the released software
> version.

No offense, but that's not really true. If I buy a product that has a big,
beautiful display with large fonts I can see across the room and I come
home and download the release version of the software and discover that,
well, those fonts are only going to show up if I (a) download a "nightly"
or (b) wait some unknown amount of time for the next release version, I
don't consider that to be a fully functional product.

I realize from a techie point of view what you're saying is basically
true, and I don't want this to sound harsh. I'm just trying to frame
things from a non-techie, end-user point of view.

Say I buy a copy of DOOM 3 because the box shows a picture with incredible
graphic detail and lighting effects (or perhaps I watched a trailer on the
DOOM 3 web site). I get the game home, install it, and discover that
although I have a super fast PC and graphics card, the version I bought
only has "pretty good" graphics, but nothing spectactular, and maybe it
only has stereo sound, not surround sound. If I want any of the above, I
have to go to the developer area of the web site and find a "nightly"
build and hope it doesn't introduce more problems than it solves. Or I
have to wait.

The question is, do you consider that version of DOOM 3 to be "fully
functional"? Certainly you can play it; at best all that you'll get with a
nightly is better graphic effects and sound, but the game play will be the
same.

But the reality is, what you bought turns out not to be what was
advertised, not unless you opt to use beta-quality software (assuming you
can even find it on the web site). Most people would feel mildly annoyed
if they found themselves in this situation -- it's not exactly false
advertising; it's certainly not "bait and switch"; but you're definitely
not getting what you thought you were getting when you bought the product.

That's a problem.

SBB

Robin Bowes
2004-09-06, 09:04
On Mon, September 6, 2004 16:40, Jack Coates said:
> ...
>
>>
>> So, slim devices are currently selling a product that only works with beta
>> software - not good.
>>
>
> Perhaps you're new to the computer business, or suffering temporary amnesia?

I'm not sure I understand your comment here.

I understood that the SBG required 5.3.0b1 to be fully functional. You replied saying
that the SBG works with 5.2.1 but:

"you only get the purty fonts with the 5.3 beta, but the
display is still entirely functional with the released software version."

I would argue that isn't fully functional - the main (only?) selling point of the new
graphical version is the "purdy" fonts.

I still maintain that, for the majority of non-technical purchasers, there should be an
official, non-beta version of slimserver that provides the full graphical experience.

>> To re-iterate my previous comments, there should be two distinct versions
>> of slimserver:
>>
>> 1. stable or release version, wich is updated to fix bugs and introduce
>> new features after they have been thoroughly tested, e.g. support for the graphic
>> display.
>
> Who's going to test these? Certainly not the people complaining about
> bugs. QA costs a lot more money than it brings in. Ditto for tech support. Yeah, that
> sucks, but that's the way things are in the computer business. This is the point where
> you'll argue that Slim Devices is entering the consumer electronics business, a bastion
> of reliability, to which I must laugh.

I'm not particularly suggesting a whole load of additional work - just better
configuration management of the release procedure. That falls squarely in Slim Devices'
court.

And laugh if you must, but for Slim Devices to survive they need to sell product in
sufficient volume, and that means the consumer market.

>> 2. development or beta version, which is where most bugs are fixed and
>> tested and where new features are first introduced and tested.
>>
>> Plus of course:
>>
>>
>> 3. CVS (or nightly builds), for those who like to be bleeding edge!
>>
>>
>> This is how a lot of (most?) software is distributed and should keep most
>> people happy.
>>
>
> examples, please. The only one I can think of is Debian.

Jack, there are any number of open source software projects that use this development
model. Have a look at http://www.sf.net and checkout how SF projects manage releases.

>
>> Non-techy users can use the latest stable version - less features but
>> stable
>
> like say, the released version, 5.2.1?
>
>> Geeks and hackers can use the latest development version - latest
>> features, but probably more bugs.
>
> Hm, you must be referring to the beta version, 5.3.0b1.
>
>
>> And masochists can update from CVS or use the latest nightly build!
>>
>>
>
> downloads/nightly/latest!
>
>> How about it?
>>
>
> If you're looking for completely separate trees, go ahead and fork, the
> code is open. If you need some help setting up a CVS server, let me know off-list. You
> might want to investigate what's already in progress though, the database-backed branch
> is not the one you want.

I'm not looking to fork anything - I'm suggesting that slimserver version management is
tightening up.

Personally, I update from CVS regularly - usually nightly - that's my choice. But I
understand and sympathise with those users of the Squeezebox that simply want to plug in
their product and listen to music without the server hanging, or the display corrupting,
or getting dropouts.

Don't get me wrong - I think slimserver and the Squeezebox are great products - but
things could be better, and I think Slim Devices management are responsive enough to
listen to constructive criticism like this.

R.
--
http://robinbowes.com

Steve Baumgarten
2004-09-06, 09:21
> Venkman: Human sacrifice, dogs and cats, living together... mass
> hysteria!

;-)

I know, none of this is the end of the world. I'm just trying to frame
things as others -- many others, but especially non-techies -- will see
it.

My dad often doesn't understand why something (generally something stupid)
is the way it is about Windows. It does very little good for me to be able
to tell him that (a) the world is not ending or (b) that here is the long
history behind why this stupid thing is the way it is, going back to the
days of DOS and CP/M before it. For people like my dad, things either work
right (i.e., the way he expects, or more generally, the way a reasonable
person would expect) or they don't.

DVDs with non-skippable FBI warnings and (in some cases) movie previews
you can only fast forward through but not skip: another perfect example.
Again, the world is not ending. On the other hand, when you bought your
DVD player you were probably under the assumption that it would work much
like your VCR -- you'd be able to watch what you wanted, how you wanted,
and not be forced to watch things you don't want to watch. It does no good
to be able to explain why this isn't so, the technical reasons behind it,
the corporate decisions behind it, etc. The reality is, it's disappointing
and a nuisance, and it's aggravating in some small way.

The studios can get away with this because they have a monopoly on the
software -- you want a copy of a Disney film, you can only get it from
Disney. So if people are slightly annoyed, what do they care?

Slimdevices can't be in that position. They have to be in the position
where you buy the product and you get something that works just as
advertised, and you're happy with it and can recommend it to your friends,
and they the product, etc.

If the out of the box experience for people is "hey, this isn't what I
paid for", that's not good. And a company like Slimdevices can't afford to
take Disney's attitude about it. (Or for that matter, yours.)

Let me also note: I think this is different than the fact that the current
product doesn't scale to 100K songs. It's reasonable to expect people to
have reasonable expectations about a product -- if that makes sense. It's
great that this is being addressed with the Info.pm rewrite, but very few
people are going to be buying the product with the expectation that
they'll be able to throw a terabyte of music at it and have it perform
well. On the other hand, I'd venture to guess that *every single person*
who's buying a SBG right now expects to be able to see those big,
beautiful fonts right out of the box. That's a reasonable expectation --
and it's unreasonable to expect those people to just deal with their
disappointment by explaining the state of the world to them or by quoting
GHOSTBUSTERS to them.

Though thank you for reminding me of that scene. What a great movie...

ftp://members.aol.com/dgresh2/ghostbusters/disaster.wav

SBB

Jack Coates
2004-09-06, 10:06
> On Mon, September 6, 2004 16:40, Jack Coates said:
>> ...
>>
>>>
>>> So, slim devices are currently selling a product that only works with
>>> beta
>>> software - not good.
>>>
>>
>> Perhaps you're new to the computer business, or suffering temporary
>> amnesia?
>
> I'm not sure I understand your comment here.

Sorry, I tend to be a little more nasty then necessary before the first
pot of coffee. Basically, I disagree with what I perceive as your point
that Slimserver is suffering from an unusual or even important number of
bugs. My experience as a consumer and a vendor in the high-tech world is
that quality is job four or five for most companies. Truly pursuing
quality drives companies out of business because they spend so much on QA
and tech support without ever reaching the goal of perfection. The time to
spend time and money on a quality issue is when sales are dropping or
about to drop because the competition is doing a better job.

Since I'm not actually seeing any of these bugs in my own installation,
I'm also not very sympathetic to the argument that it's thoroughly broken
and unusable.
--
Jack At Monkeynoodle.Org: It's A Scientific Venture...
"Believe what you're told; there'd be chaos if everyone thought for
themselves." -- Top Dog hotdog stand, Berkeley, CA

Robin Bowes
2004-09-06, 10:26
On Mon, September 6, 2004 18:06, Jack Coates said:
>
> Sorry, I tend to be a little more nasty then necessary before the first
> pot of coffee. Basically, I disagree with what I perceive as your point that Slimserver
> is suffering from an unusual or even important number of bugs.

Then your perception of my point is incorrect.

My point, if I had to narrow my previous posts down to just one point, is that the
software required to fully support the new graphical squeezebox is still in beta.

The way the software is currently managed and released is not optimal. This is not the
same as saying that "slimserver is suffering from an unusual or even important number of
bugs".

In fact, your statement illustrates the problem perfectly. What do you mean by
"slimserver"? Do you mean v5.2.1? Or v5.3.0b1? Or the nightly from 2004-08-27? Or the
latest CVS version? Get my point?

IMHO, Slim Devices should formalise the configuration management of slimserver and make
official releases of both stable and development branches of the code. By introducing a
clearly defined lifecycle into the development process, the developers can focus on
fixing bugs prior to a new release, or adding and testing a clearly defined set of
features.

Does that better explain my point?

> My experience as a
> consumer and a vendor in the high-tech world is that quality is job four or five for
> most companies. Truly pursuing quality drives companies out of business because they
> spend so much on QA and tech support without ever reaching the goal of perfection. The
> time to spend time and money on a quality issue is when sales are dropping or about to
> drop because the competition is doing a better job.

Nobody's is talking about perfection - quality is not a binary property!

>
> Since I'm not actually seeing any of these bugs in my own installation,
> I'm also not very sympathetic to the argument that it's thoroughly broken
> and unusable.

And who, pray tell, is making that argument?

R.
--
http://robinbowes.com

Mike Kozlowski
2004-09-06, 10:53
On Mon, 6 Sep 2004, Jack Coates wrote:

>> So, slim devices are currently selling a product that only works with beta
>> software -
>> not good.
>
> Perhaps you're new to the computer business, or suffering temporary amnesia?

On the contrary, most significant open source (and, for that matter,
commercial) projects work in just the way he mentioned. You branch for a
release version, continue development in a dev branch, and back-port
important bug fixes to release.

> examples, please. The only one I can think of is Debian.

Mozilla. The Linux kernel. Perl. Python. Windows. Office. PaintShop
Pro. I could go on for a while.

>> Non-techy users can use the latest stable version - less features but
>> stable
>
> like say, the released version, 5.2.1?

Yes. The problem, though, is that there are some serious bugs with 5.2.1
-- theme switching can leave you with a broken UI and no apparent way to
switch back, for instance -- that have been fixed in the dev branch, but
not back-ported to the stable version. So when someone says, "I'm
encountering this serious bug that makes the application unusable for me,"
the response is, as often as not, "Upgrade to this untested, unstable
version that will fix your bug, but will probably introduce new ones."

Predictably, the user then keeps encountering bug after bug, and has to
keep upgrading to a stream of unstable versions to get bugfixes;
eventually, they decide that this whole SlimServer thing is just too
shoddy for words and not ready for primetime yet.

If, instead, fixes for serious bugs were backported onto 5.2.x, users
could just download the latest stable release and go on with their lives,
using a rock-solid, production-ready version of SlimServer.

Porting bug fixes to an old stable version isn't always easy and Slim may
believe they can't afford the overhead to do that; but the increase in
customer satisfaction would more than make it worth it, in my opinion.

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

Daryle A. Tilroe
2004-09-06, 18:53
Mike Kozlowski wrote:

> Yes. The problem, though, is that there are some serious bugs with
> 5.2.1 -- theme switching can leave you with a broken UI and no apparent
> way to switch back, for instance -- that have been fixed in the dev
> branch, but not back-ported to the stable version. So when someone
> says, "I'm encountering this serious bug that makes the application
> unusable for me," the response is, as often as not, "Upgrade to this
> untested, unstable version that will fix your bug, but will probably
> introduce new ones."

I think this is exactly how I would characterize the situation.
IMHO since I got into the squeezebox last winter/spring there
has yet to be a rock solid release of the software and firmware.
Don't get me wrong; personally I don't particularly mind being
near the bleeding edge to get a new feature but we really lack a
baseline that I think existed for the older SLIMP3 about a year
or so ago. I think release versions need to be better nailed
down and then fixed for major bugs even as development goes on.
To many people are running to many different versions of the
nightlies just to fix a bug or two that probably should be
back ported. Another consideration is that the nightlies should
become selected, near stable, release points on a less frequent
basis (say on or two a month). If you really want the most
recent there is always the CVS. This would allow better testing
by more people running the same version. For example right now
I'm really happy with 2004-08-31; so much so that I don't think
I'm going to touch anything else until a fixed beta point comes
up. But how many people are running that version?

--
Daryle A. Tilroe