Re: [OSM-dev] helping #switch2osm

[Roger Weeks]

Hello

I'm the person Mikel was helping (thanks Mikel!) and I joined the list to
give my thoughts to this discussion. I think you have a couple of different
types of people you may want to think about here.

1. IT/devops

I fall into this category. The project I'm working on will require us to
have our own tile server that takes planet_osm, processes it and produces
updates to our client devices. Other people that fall into this category
will be anyone who wants to build their own GIS database from planet_osm,
for whatever reason. They'll also want to generate and serve tiles.

My major problem all through this process has been the vast amounts of
helpful and conflicting information. You'll note that the process I was
following to get a working PostGIS database populated with planet_osm, was
compiled by someone else and was not hosted at OSM. That process also is
flawed and needed a bunch of other things to make it work.

My suggestion for either a chef recipe, a juju charm, a puppet script, or
 was meant to fix this
situation. If someone wants to set up a PostGIS OSM database, it should be
a repeatable process that is easy to do and scalable across multiple *nix
platforms. (I won't address Windows here as that's a whole other can of
worms).

Taking that idea a step further, if we were to provide a canned VM in KVM
or Xen formats that provides not just the host OS, but preconfigured
Postgres, PostGIS, an imported planet_osm, and then perhaps a choice of
mapnik, mod_tile, tilestache, or any of the other plethora of tile
generation and serving options, that would be an IDEAL way to summon
IT/devops folks who want to build their own map server infrastructure.

The last thing this needs is some really good documentation and
recommendations for WHAT to do. There are enough options out there
especially for tile generation and serving, that there needs to be a good
discussion of what you would want to use and why.

2. Maps developers and users

I'm not as familiar with the needs here but I think you are going to want
to do things like what MapBox is doing with Tilemill and what Cloudmade is
doing. The same comments I made above about recommendations of what you
want to use and why need to be easily accessible.

Happy to continue this discussion. I would be interested in participating
in development of both recipes and VMs.

Roger Weeks

On Jan 13, 2012 3:37 AM, "Mikel Maron"  wrote:

> > Hi
> >
> > Over the past couple days, I gave some assistance to someone setting up
> their own tile infrastructure. They mostly got along on their own,
> following http://weait.com/content/build-your-own-openstreetmap-server,
> and picking through the wiki, but they definitely could've had an easier
> time.
> >
> > And we should expect more of this soon, and we want this! Nestoria and
> StreetEasy switching, the Wired article. Just the beginning.
>
> Yes. The time when people are recognising the power of OSM is here. We
> should welcome everyone.
>
> > How do we help folks get their tile servers up and running even easier.
> Here's some suggestions we discussed...
> >
> > 1) Clear instructions, at an easily findable address (outside the wiki,
> which is hard to follow) (anyone register switch2osm.com? or tryosm.org).
> Should include some description of the different options, with links for
> getting into the details ... but at least a clear guide through the well
> worn path of setting up a tile server.
> >
> > 2) A packaged solution, like a chef recipe, to install everything needed.
> >
> > 3) A VM image for an open source virtualization package, that has
> everything installed. May not necessarily be for production, but at least
> make it easy to try things out.
> >
> > 4) 
> >
> Might be an overkilling idea, why not some quick screen casts? I guess
> videos will be more easier to grasp than guides.
>
> > Would the EWG be the place to get this together?
> >
> > Cheers
> > Mikel
> >
> >
> > == Mikel Maron ==
> > +14152835207 @mikel s:mikelmaron
>
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] helping #switch2osm

[Sajjad Anwar]

On Jan 13, 2012 3:37 AM, "Mikel Maron"  wrote:
>
> Hi
>
> Over the past couple days, I gave some assistance to someone setting up
their own tile infrastructure. They mostly got along on their own,
following http://weait.com/content/build-your-own-openstreetmap-server, and
picking through the wiki, but they definitely could've had an easier time.
>
> And we should expect more of this soon, and we want this! Nestoria and
StreetEasy switching, the Wired article. Just the beginning.

Yes. The time when people are recognising the power of OSM is here. We
should welcome everyone.

> How do we help folks get their tile servers up and running even easier.
Here's some suggestions we discussed...
>
> 1) Clear instructions, at an easily findable address (outside the wiki,
which is hard to follow) (anyone register switch2osm.com? or tryosm.org).
Should include some description of the different options, with links for
getting into the details ... but at least a clear guide through the well
worn path of setting up a tile server.
>
> 2) A packaged solution, like a chef recipe, to install everything needed.
>
> 3) A VM image for an open source virtualization package, that has
everything installed. May not necessarily be for production, but at least
make it easy to try things out.
>
> 4) 
>
Might be an overkilling idea, why not some quick screen casts? I guess
videos will be more easier to grasp than guides.

> Would the EWG be the place to get this together?
>
> Cheers
> Mikel
>
>
> == Mikel Maron ==
> +14152835207 @mikel s:mikelmaron
>
> ___
> dev mailing list
> dev@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/dev
>
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] helping #switch2osm

[Mike N]

On 1/12/2012 5:06 PM, Mikel Maron wrote:


2) A packaged solution, like a chef recipe, to install everything needed.



 FYI - there was this project on last Summer of Code:

http://wiki.openstreetmap.org/wiki/Fast_Track_Deployment_of_Customised_OSM_Tile_Server

 I'm not sure what all would be usable from that.

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] 23 ideas for osm and its forks

[Mike Dupont]

On Fri, Jan 13, 2012 at 2:19 AM, Gregory 
 wrote:

> Though I don't watch the mailing list much, this looks like a troll's list.


I have tried to summarize the ideas that I have been working on in the past
three years.
Lots of though put into this. This is not meant to be a flame or troll
thing, but an honest attempt to bring fresh ideas to the table.

e.g. "host the CC-by-sa data that will be deleted". Although I'm pretty
sure the OSMF's plan for ODbL switch makes it clear the CC-by-sa data will
not be deleted and will remain available.
I will comment on that point in the ideascale and explain it.

I got bored after that.
>
> that is a shame.
I will have to put more work into this.
mike


> On 13 January 2012 01:00, Mike Dupont wrote:
>
>> Hello world,
>>
>> I have written down all my ideas so far on an ideascale i setup
>>
>>http://fosm.ideascale.com/
>>
>> here you have 23 new ideas for #osm,
>> independent of how they are implemented, all applicable to osm commonmap
>> fosm etc.
>> you are invited to join the conversation.
>>
>> lets make the world a better place
>>
>> happy hacking,
>>
>> mike
>>
>> --
>> James Michael DuPont
>> Member of Free Libre Open Source Software Kosova http://flossk.org
>>
>> ___
>> dev mailing list
>> dev@openstreetmap.org
>> http://lists.openstreetmap.org/listinfo/dev
>>
>>
>
>
> --
> Gregory
> o...@livingwithdragons.com
> http://www.livingwithdragons.com
>



-- 
James Michael DuPont
Member of Free Libre Open Source Software Kosova http://flossk.org
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] 23 ideas for osm and its forks

[Gregory]

Though I don't watch the mailing list much, this looks like a troll's list.
e.g. "host the CC-by-sa data that will be deleted". Although I'm pretty
sure the OSMF's plan for ODbL switch makes it clear the CC-by-sa data will
not be deleted and will remain available.

I got bored after that.

On 13 January 2012 01:00, Mike Dupont wrote:

> Hello world,
>
> I have written down all my ideas so far on an ideascale i setup
>
>http://fosm.ideascale.com/
>
> here you have 23 new ideas for #osm,
> independent of how they are implemented, all applicable to osm commonmap
> fosm etc.
> you are invited to join the conversation.
>
> lets make the world a better place
>
> happy hacking,
>
> mike
>
> --
> James Michael DuPont
> Member of Free Libre Open Source Software Kosova http://flossk.org
>
> ___
> dev mailing list
> dev@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/dev
>
>


-- 
Gregory
o...@livingwithdragons.com
http://www.livingwithdragons.com
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

[OSM-dev] 23 ideas for osm and its forks

[Mike Dupont]

Hello world,

I have written down all my ideas so far on an ideascale i setup

   http://fosm.ideascale.com/

here you have 23 new ideas for #osm,
independent of how they are implemented, all applicable to osm commonmap
fosm etc.
you are invited to join the conversation.

lets make the world a better place

happy hacking,

mike

-- 
James Michael DuPont
Member of Free Libre Open Source Software Kosova http://flossk.org
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] helping #switch2osm

[Matthias Meißer]

Am 12.01.2012 23:37, schrieb Matthias Meißer:

Am 12.01.2012 23:06, schrieb Mikel Maron:

4) 

Get easier in contact with local communities/multiplyer/"evangelists"?
But promoting this contacts is quiet hard... Sometimes somebody asks on
twitter, sometimes it's only a discussion within the company team :/


maybe what will help for commercial partners is something like that:
https://partner.microsoft.com/UK/productssolutions/40174347

So a wikipage (sayhttp://wiki.openstreetmap.org/wiki/switch2osm) with 
short facts/links on the aspects that are covered by a change:

-technical
-legal/financial
-communication/best practise
This sounds to me good, as we have so much options (what 
webmap-framework, what tile provider, what styling, ...) this will give 
companies a helping hand for a quickstart, doesn't it?


bye
Matthias
(user:!i!)

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] helping #switch2osm

[Simon Poole]

Kais Ubuntu tile server package is really easy to install (one serious 
caveat there) I don't think it will get a lot simpler than that even 
with substantially more work.


The install instructions are simple and work (again that caveat), what 
is a bit lacking is post-install documentation (essentially mod_tile, 
renderd and osm2pgsql).


The big caveat I have is not specific to Kais package, it is the general 
issue that we have no release engineering/testing for a  number of  our 
tools. This bites particularly hard when something is supposed to be 
plug and play and you are not aware that you are essentially just 
running the latest development version with a couple of changes that 
might work.


If we really want to make things easier for third parties we are not 
going to get around having substantially more discipline in our software 
release procedures (and having such in the first place).


Simon

Am 12.01.2012 23:06, schrieb Mikel Maron:

Hi

Over the past couple days, I gave some assistance to someone setting 
up their own tile infrastructure. They mostly got along on their own, 
following 
http://weait.com/content/build-your-own-openstreetmap-server, and 
picking through the wiki, but they definitely could've had an easier time.


And we should expect more of this soon, and we want this! Nestoria and 
StreetEasy switching, the Wired article. Just the beginning.
How do we help folks get their tile servers up and running even 
easier. Here's some suggestions we discussed...


1) Clear instructions, at an easily findable address (outside the 
wiki, which is hard to follow) (anyone register switch2osm.com? or 
tryosm.org). Should include some description of the different options, 
with links for getting into the details ... but at least a clear guide 
through the well worn path of setting up a tile server.


2) A packaged solution, like a chef recipe, to install everything needed.

3) A VM image for an open source virtualization package, that has 
everything installed. May not necessarily be for production, but at 
least make it easy to try things out.


4) 

Would the EWG be the place to get this together?

Cheers
Mikel


== Mikel Maron ==
+14152835207 @mikel s:mikelmaron


___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] helping #switch2osm

[Matthias Meißer]

Am 12.01.2012 23:06, schrieb Mikel Maron:

Hi

Over the past couple days, I gave some assistance to someone setting up
their own tile infrastructure. They mostly got along on their own,
following http://weait.com/content/build-your-own-openstreetmap-server,
and picking through the wiki, but they definitely could've had an easier
time.

And we should expect more of this soon, and we want this! Nestoria and
StreetEasy switching, the Wired article. Just the beginning.
How do we help folks get their tile servers up and running even easier.
Here's some suggestions we discussed...

1) Clear instructions, at an easily findable address (outside the wiki,
which is hard to follow) (anyone register switch2osm.com? or
tryosm.org). Should include some description of the different options,
with links for getting into the details ... but at least a clear guide
through the well worn path of setting up a tile server.

Short domains are welcome, similar to getfirefox.org etc. :)
If you have any suggestions on specific pages or if we should create 
some easier redirects within the wiki, just give us a ping.


Might sound curious, but what most people looking for, seems to be a 
well structured book?

http://wiki.openstreetmap.org/wiki/Books


2) A packaged solution, like a chef recipe, to install everything needed.

http://wiki.openstreetmap.org/wiki/Ubuntu_tile_server is on the way.


3) A VM image for an open source virtualization package, that has
everything installed. May not necessarily be for production, but at
least make it easy to try things out.
That was during the discussion of the ubuntu package, too. Most expect a 
VM to be to slow for a good tile service.


4) 
Get easier in contact with local communities/multiplyer/"evangelists"? 
But promoting this contacts is quiet hard... Sometimes somebody asks on 
twitter, sometimes it's only a discussion within the company team :/


bye
Matthias



Would the EWG be the place to get this together?

Cheers
Mikel


== Mikel Maron ==
+14152835207 @mikel s:mikelmaron



___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev



___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] helping #switch2osm

[Ian Dees]

During the book sprint last year I put together Richard's blog post and
some stuff related to renderd/tirex to create this chapter in the book:

http://en.flossmanuals.net/openstreetmap/setting-up-your-own-tile-server/

It should be fairly complete and copy-pasteable. I'd be more than happy to
make changes if anyone sees errors.

On Thu, Jan 12, 2012 at 4:06 PM, Mikel Maron  wrote:

> Hi
>
> Over the past couple days, I gave some assistance to someone setting up
> their own tile infrastructure. They mostly got along on their own,
> following http://weait.com/content/build-your-own-openstreetmap-server,
> and picking through the wiki, but they definitely could've had an easier
> time.
>
> And we should expect more of this soon, and we want this! Nestoria and
> StreetEasy switching, the Wired article. Just the beginning.
> How do we help folks get their tile servers up and running even easier.
> Here's some suggestions we discussed...
>
> 1) Clear instructions, at an easily findable address (outside the wiki,
> which is hard to follow) (anyone register switch2osm.com? or tryosm.org).
> Should include some description of the different options, with links for
> getting into the details ... but at least a clear guide through the well
> worn path of setting up a tile server.
>
> 2) A packaged solution, like a chef recipe, to install everything needed.
>
> 3) A VM image for an open source virtualization package, that has
> everything installed. May not necessarily be for production, but at least
> make it easy to try things out.
>
> 4) 
>
> Would the EWG be the place to get this together?
>
> Cheers
> Mikel
>
>
> == Mikel Maron ==
> +14152835207 @mikel s:mikelmaron
>
> ___
> dev mailing list
> dev@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/dev
>
>
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

[OSM-dev] helping #switch2osm

[Mikel Maron]

Hi

Over the past couple days, I gave some assistance to someone setting up their 
own tile infrastructure. They mostly got along on their own, 
following http://weait.com/content/build-your-own-openstreetmap-server, and 
picking through the wiki, but they definitely could've had an easier time.


And we should expect more of this soon, and we want this! Nestoria and 
StreetEasy switching, the Wired article. Just the beginning. 
How do we help folks get their tile servers up and running even easier. Here's 
some suggestions we discussed...

1) Clear instructions, at an easily findable address (outside the wiki, which 
is hard to follow) (anyone register switch2osm.com? or tryosm.org). Should 
include some description of the different options, with links for getting into 
the details ... but at least a clear guide through the well worn path of 
setting up a tile server.

2) A packaged solution, like a chef recipe, to install everything needed.

3) A VM image for an open source virtualization package, that has everything 
installed. May not necessarily be for production, but at least make it easy to 
try things out.

4) 

Would the EWG be the place to get this together?

Cheers
Mikel


== Mikel Maron ==
+14152835207 @mikel s:mikelmaron___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [josm-dev] Dividing houses

[David Groom]

Never tried this plugin, but it might do what you want 
http://wiki.openstreetmap.org/wiki/JOSM/Plugins/Terracer

David
 
  - Original Message - 
  From: Oliver Raupach 
  To: josm-...@openstreetmap.org 
  Sent: Thursday, January 12, 2012 10:47 AM
  Subject: [josm-dev] Dividing houses


  Hi,

  I have the following technical question: Houses were often drawn as a 
  big long block. In reality, there are however a number of individual 
  houses. Is there a way to quickly cut up these large blocks into small 
  blocks? Right now my workflow is a bit more complicated here. I add a 
  new point - split the line - draw the dividing line - delete the help 
  points. Or I draw everything new

  --

 Oliver Raupach

  ___
  josm-dev mailing list
  josm-...@openstreetmap.org
  http://lists.openstreetmap.org/listinfo/josm-dev
___
josm-dev mailing list
josm-...@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev

Re: [OSM-dev] mod_tile performance enhancements?

[Frederik Ramm]

Hi,

On 01/12/2012 05:18 PM, Jason Lee wrote:

The experience is rather like WMS vs TMS where you frequently see a
blank map area in WMS. Bearing in mind I am using NUM_THREADS = 32 which
means there are more processors to deal with the higher number of
Metatile(2) requests.


That will help you very little unless you have a disk with about 32 
spindles ;)


Bye
Frederik

--
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] mod_tile performance enhancements?

[Dane Springmeyer]

I'll add:

1) As you likely know png/zlib compression can be quite slow (and sometimes 
constitutes a meaningful % of rendering time), so if you are running >= Mapnik 
2.0.0 you can use a format string like "png:z=1" to increase encoding time at 
the slight cost of size. [1]

2) In TileMill we default to 2X metatiles for the usability reason you cite 
below, but of course its configurable like mod_tile.

3) In Mapnik in current trunk we're currently refactoring core storage of 
features to make them smaller and faster - this should benefit performance of 
very large metatiles that pull a big payload. [2]

Dane

[1] https://github.com/mapnik/mapnik/wiki/OutputFormats
[2] https://github.com/mapnik/mapnik/issues/834


On Jan 12, 2012, at 8:18 AM, Jason Lee wrote:

> That's a good point, thanks. But from a User Experience point of view - 
> getting lots of tiles returned in 2's within milliseconds is better than 
> waiting 10 seconds and then getting a whole bunch of tiles which, once 
> generated, the user has a better experience panning around - until they reach 
> another new area to render tiles.
> 
> The experience is rather like WMS vs TMS where you frequently see a blank map 
> area in WMS. Bearing in mind I am using NUM_THREADS = 32 which means there 
> are more processors to deal with the higher number of Metatile(2) requests.
> 
> Jason
> 
> 
> On Thu, Jan 12, 2012 at 2:56 PM, Frederik Ramm  wrote:
> Hi,
> 
> 
> On 01/12/12 14:32, Jason Lee wrote:
> I am focusing and playing around with the METATILE parameter and get
> performance of 1 tile request :-
> 
> METATILE(16) - 10 secs
> METATILE(8) - 3.5 secs
> METATILE(4) - 1.5 secs
> METATILE(2) - 0.6 secs
> 
> Maybe you are misunderstanding what this means. This means effectively
> 
> METATILE(16) - 10 secs - 256 tiles - 0.039 sec per tile
> METATILE(8) - 3.5 secs -  64 tiles - 0.054 sec per tile
> METATILE(4) - 1.5 secs -  16 tiles - 0.094 sec per tile
> METATILE(2) - 0.6 secs -   4 tiles - 0.150 sec per tile
> 
> 
> So it seems the best performance for rendering a NEW TILE is to use the
> minimum metatile size of 2 (default is 8).
> 
> This is true if you want to render *one tile only*. If however you have an 
> user behind an OpenLayers client who is likely to requires several 
> neighbouring tiles, or if you intend to render a large area anyway, then the 
> larger METATILE numbers give you much better performance.
> 
> Bye
> Frederik
> 
> 
> ___
> dev mailing list
> dev@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/dev
> 
> ___
> dev mailing list
> dev@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/dev

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] mod_tile performance enhancements?

[Jason Lee]

That's a good point, thanks. But from a User Experience point of view -
getting lots of tiles returned in 2's within milliseconds is better than
waiting 10 seconds and then getting a whole bunch of tiles which, once
generated, the user has a better experience panning around - until they
reach another new area to render tiles.

The experience is rather like WMS vs TMS where you frequently see a blank
map area in WMS. Bearing in mind I am using NUM_THREADS = 32 which means
there are more processors to deal with the higher number of Metatile(2)
requests.

Jason


On Thu, Jan 12, 2012 at 2:56 PM, Frederik Ramm  wrote:

> Hi,
>
>
> On 01/12/12 14:32, Jason Lee wrote:
>
>> I am focusing and playing around with the METATILE parameter and get
>> performance of 1 tile request :-
>>
>> METATILE(16) - 10 secs
>> METATILE(8) - 3.5 secs
>> METATILE(4) - 1.5 secs
>> METATILE(2) - 0.6 secs
>>
>
> Maybe you are misunderstanding what this means. This means effectively
>
> METATILE(16) - 10 secs - 256 tiles - 0.039 sec per tile
> METATILE(8) - 3.5 secs -  64 tiles - 0.054 sec per tile
> METATILE(4) - 1.5 secs -  16 tiles - 0.094 sec per tile
> METATILE(2) - 0.6 secs -   4 tiles - 0.150 sec per tile
>
>
>  So it seems the best performance for rendering a NEW TILE is to use the
>> minimum metatile size of 2 (default is 8).
>>
>
> This is true if you want to render *one tile only*. If however you have an
> user behind an OpenLayers client who is likely to requires several
> neighbouring tiles, or if you intend to render a large area anyway, then
> the larger METATILE numbers give you much better performance.
>
> Bye
> Frederik
>
>
> __**_
> dev mailing list
> dev@openstreetmap.org
> http://lists.openstreetmap.**org/listinfo/dev
>
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

[josm-dev] tab not working any more?

[Martin Koppenhoefer]

In the latest version the "TAB"-key doesn't seem to work any more (at
least not to change the column/row) on my system (ubuntu 11.04, sun
java). Am I the only one or is this due to some recent updates?

cheers,
Martin

___
josm-dev mailing list
josm-...@openstreetmap.org
http://lists.openstreetmap.org/listinfo/josm-dev

Re: [OSM-dev] mod_tile performance enhancements?

[Frederik Ramm]

Hi,

On 01/12/12 14:32, Jason Lee wrote:

I am focusing and playing around with the METATILE parameter and get
performance of 1 tile request :-

METATILE(16) - 10 secs
METATILE(8) - 3.5 secs
METATILE(4) - 1.5 secs
METATILE(2) - 0.6 secs


Maybe you are misunderstanding what this means. This means effectively

METATILE(16) - 10 secs - 256 tiles - 0.039 sec per tile
METATILE(8) - 3.5 secs -  64 tiles - 0.054 sec per tile
METATILE(4) - 1.5 secs -  16 tiles - 0.094 sec per tile
METATILE(2) - 0.6 secs -   4 tiles - 0.150 sec per tile


So it seems the best performance for rendering a NEW TILE is to use the
minimum metatile size of 2 (default is 8).


This is true if you want to render *one tile only*. If however you have 
an user behind an OpenLayers client who is likely to requires several 
neighbouring tiles, or if you intend to render a large area anyway, then 
the larger METATILE numbers give you much better performance.


Bye
Frederik

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] mod_tile performance enhancements?

[Jason Lee]

I am focusing and playing around with the METATILE parameter and get
performance of 1 tile request :-

METATILE(16) - 10 secs
METATILE(8) - 3.5 secs
METATILE(4) - 1.5 secs
METATILE(2) - 0.6 secs

So it seems the best performance for rendering a NEW TILE is to use the
minimum metatile size of 2 (default is 8). But, once it is cached, does
that mean a Metatile size of 2 is not as fast as a Metatile of 16 when
fetching a CACHED TILE or just that the disk storage is not so efficient?


On Thu, Jan 12, 2012 at 11:49 AM, Martin Koppenhoefer <
dieterdre...@gmail.com> wrote:

> 2012/1/12 Jason Lee :
> > Thanks for the tip Frederik,
> > I enabled Postgres logging and the SQL selects are quite fast (under
> > 1000ms). I am guessing the load maybe on the tile rendering and tile
> cutting
> > maybe.
>
>
> You will have to add the time of those single requests (for every
> db-based mapnik-layer a request is issued) to get the total selection
> time for a (meta) tile.
>
> cheers,
> Martin
>
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

[OSM-dev] Nominatim country limits problems

[Adrian Chapela]

Hello,

I am using Nominatim to translate positions into address. I have load
Colombia country and now I am testing it. The problem is that I have
found positions which are in Venezuela and Nominatim is translating as
Colombia, one example is
/reverse?format=xml&lat=10&lon=-67.2&zoom=18&addressdetails=1 . How
could I modify this in Openstreetmaps database to solve that ?  I
haven't loaded Venezuela to my server but I am not interesting on
translating (by now) Venezuela positions but I don't want to translate a
position that it's outside Colombia as it is inside.

Thank you.

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] mod_tile performance enhancements?

[Martin Koppenhoefer]

2012/1/12 Jason Lee :
> Thanks for the tip Frederik,
> I enabled Postgres logging and the SQL selects are quite fast (under
> 1000ms). I am guessing the load maybe on the tile rendering and tile cutting
> maybe.


You will have to add the time of those single requests (for every
db-based mapnik-layer a request is issued) to get the total selection
time for a (meta) tile.

cheers,
Martin

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] mod_tile performance enhancements?

[Jason Lee]

Thanks for the tip Frederik,
I enabled Postgres logging and the SQL selects are quite fast (under
1000ms). I am guessing the load maybe on the tile rendering and tile
cutting maybe.

On Wed, Jan 11, 2012 at 11:43 AM, Frederik Ramm  wrote:

> Hi,
>
>
> On 01/11/2012 11:39 AM, Jason Lee wrote:
>
>> Ten seconds does sound quite slow and I was wondering what can be done
>> to improve it. Just out of curiosity, I tried disabling all the 7 Layers
>> except for 1 Layer in the Mapnik stylesheet, so the rendering requires a
>> lot less db querying and map rendering
>>
>
> To get a definitve answer on how much time is spent in database
> processing, set log_min_duration_statement to 0 (default is -1 = disabled)
> in your postgresql.conf, kill -HUP the database, and re-run your request.
> You'll get durations for every single statement. It is very easy to
> overlook a performance killer here. If you find a statement that takes
> seconds, run an "explain" on it to make sure it uses an index, or create an
> appropriate index if it doesn't.
>
> Bye
> Frederik
>
> --
> Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"
>
>
> __**_
> dev mailing list
> dev@openstreetmap.org
> http://lists.openstreetmap.**org/listinfo/dev
>
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev