The AWS SDKs take care of getting the keys made available to the EC2
instance via IAM Roles. But if you are running something custom you can get
the key info using a meta data call. See this doc page.
This method is specific to Amazon EC2, but there is a gist located here
https://gist.github.com/mwkorver/1ef45abac3871360f2b1
that is an example of how to deploy a cluster of Mapservers using a remote
map file AND data. Both map file and source geotiffs are accessed via HTTP
because they both sit
to me.
-Mark Korver
On Wed, Oct 28, 2015 at 2:31 AM, Rahkonen Jukka (MML) <
jukka.rahko...@maanmittauslaitos.fi> wrote:
> Hi,
>
>
>
> I wrote last year
> https://lists.osgeo.org/pipermail/mapserver-users/2014-May/076442.html
> and asked if it could be possible to up
First of all, very nice work on MS4W. Maybe the next time I give a demo for
AWS, I will flip from my Ubuntu based instances running on EC2 to Windows.
In the interest of getting more people access to all of this effort,
especially new users, I am wondering how much interest there is out there
in
In order to serve a lot of requests you generally need to cache. That would
be true of any query against a db, or against a shapefile. The point is,
you don't want to go to db or to disk.
In order make this happen for maps the usual pattern is to serve tiles. If
your data gets updated every 5
Hi Experts,
I need a little advice on how to do lots of aerial image GeoTiffs quickly.
I have a large region with many GeoTiffs covering 9 UTM zones. Typically I
would put together a map file with layers per utm zone and use shapefile
index to point at the GeoTiffs. I know that declaring the
Jukka, thank you for the reminder on -t_srs, I had forgotten about that.
It makes sense if using the shapefile index to build that in the srs you
plan to output. Now that I have been reminded of that feature of gdaltindex
I can see that would be faster than parsing a large vrt file.
Thanks.
On
Hi, This is more a gdal question than Mapserver, but I figured others have
bumped into this problem already. My goal is to get this data working in
Mapserver
I am trying to work with NAIP 2012 jp2 files downloaded from the tnris.orgsite.
You can see the announcement here. I am working with the
Strictly speaking you don't have to re-project all of the data. Mapserver
can resample rasters on the fly into new projections. You can build a map
file that references your different UTM zones as layers and serves up one
WMS response as a jpeg image etc.
However, this could be very slow
You need to upload your map file for anybody to be able to help on this.
On Mon, Mar 19, 2012 at 5:13 AM, Mark Volz volzm...@yahoo.com wrote:
Hello,
I am using mapserver to consume a wms air photo service, and reproject the
image so that it can be used with a program that requires a wms
How you approach this depends on a number of variables.
Because rasters tend to stress both your server resources and network
bandwidth to client, how far you go in optimizing your system is a
function of the expected amount of data access ( # of users X
frequency of access ), the expectations of
Yes, MapServer is all about making appropriately scaled maps on the
fly, and lot more. To get started you need to go through a tutorial
on MapServer, how to create map files, get base image layer working
etc.
I am not up to date on what good newbie tutorials are out there.
Maybe somebody could
Assuming you are using something like Openlayers as the client you
could use json to return the data pts formatted as a linestring. This
would give you better options to show attribute data as a popup etc.
like in the example
http://openlayers.org/dev/examples/vector-formats.html
On Wed, Jun
there is also code here
http://www.maptiler.org/google-maps-coordinates-tile-bounds-projection/
On Fri, Jun 10, 2011 at 3:35 PM, Armin Burger armin.bur...@gmx.net wrote:
I discovered now the conversion method in the Mapserver source code in
maptile.c = msTileSetExtent().
armin
On
As Steve noted, if the input is TIF, then you prob want to be using
gdaltindex rather than tile4ms.
Now we want to put them together (the Tiles),
say, first, all the tiles from the images of the same province,
and later, all the tiles from the provinces of the same projection.
when
aproach we deliver fast images with low resolution
and only the pertinent pieces in an higher resolution...
Thanks for asking, any ideas?
-JC
-Mensaje original-
De: Mark Korver [mailto:mwkor...@gmail.com]
Enviado el: Jueves, 09 de Junio de 2011 12:10 p.m.
Para: Juan Carlos Fuchs
I would think that the thinking here is that a server only scales well
to some number of concurrent requests. That beyond that number, all
requests suffer. In this case it's better to queue the requests
rather than try to process them, meaning queuing improves performance
overall.
On Thu, May
1. Can mapserver find the nearest point directly? If so, how do I set
it up?
normally this would be PostGIS query as in
http://osgeo-org.1803224.n2.nabble.com/get-the-closest-feature-in-a-point-query-td5819760.html
2. Is this faster than finding the polygon the click is inside?
I would
I agree with Andreas,
if you just want to manipulate the shapefile and continue to use a
moded shapefile with MapServer then Spatialite ( or ogr2ogr ) can help
you, but if you want query flexibility with MapServer then postGIS is
the way to go.
On Wed, May 4, 2011 at 3:23 AM, Eichner, Andreas -
Spatialite jumps into my mind.
http://www.gaia-gis.it/spatialite/
On Tue, May 3, 2011 at 3:19 PM, Varun saraf vsaraf@gmail.com wrote:
Hi,
Thanks a lot Andreas. That JOIN was the culprit as you rightly
suggested. Once I removed the join, the performance increased
exponentially. The 5
Great answer from Jukka, just adding abit.
With jpg you have to have a world file. One work-around to that is to
use tiff or more specifically Gtif/geotif which internally contains
the world file/projection info so doesn't need a external world file.
As Jukka explains the gtif way is better
Maybe you can just define the layer 2X in the mapfile. Put the
altered layer on top of the regular layer? I am assuming that you can
get the good values to be transparent and not white.
On Tue, Apr 26, 2011 at 11:57 AM, Matt Bartolome mattxb...@gmail.com wrote:
Hello,
I have a funky raster I'm
if you purpose is to replace the google base layer with something
different this might help
http://wiki.openstreetmap.org/wiki/Google_Maps_Example#Example_Using_Google_Maps_API_V3
that allows you to use OSM as basemap.
To use MapServer behind this rather than just getting from OSM, or
CloudMade
GlobalMapper has been updated in ver 12 to work with tiled services,
but I have not tried WMS-C with it.
MapServer will also work with tiled services from gdal ver 1.7+ see
http://www.gdal.org/frmt_wms.html
Another angle on this is to put MapProxy in front of it which is easy
way to wms a tms
You need to be using png with transparency, not jpeg. Jpeg does not
support transparency.
format=image/png
also
transparent=true
quick test is:
If you are merging tiles together to create 1 large tif, why would you
need to create an index to it? I would think you would just directly
reference the tif.
so, rather than
TILEINDEX ../data/myindex
skip the index building steps
just
DATA ../data/big.tif
On Fri, Feb 25, 2011 at 9:08 AM,
On Thu, Feb 3, 2011 at 1:06 AM, Oliver Tonnhofer o...@omniscale.de wrote:
Hi Greg,
On 02.02.2011, at 22:45, gregcorradini wrote:
It sounds like I can use GDAL_WMS (example here
http://www.gdal.org/frmt_wms_openstreetmap_tms.xml) to talk to a TMS cache
or TileCache directly and get back
There are ways to use S3 as the store for source images by using tools
like s3fs (FUSE-based file system backed by Amazon S3) and writing
front end code the intercepts the incoming WMS request, filters using
a grid, then routes to the appropriate EC2 MapServer instance. This
allows particular
I think current S3 pricing for 50TB at 0.125/GB comes to about
6250/month. That is for Designed for 99.9% Durability. I
can't even count that many 9s. I know you can buy HDs for about
$50/TB = $2500 for that 50TB. Assuming you replace them all once a
year its still only $5000.
But we
for the required
redundancy if you can call it that.
On 1/29/2011 12:14 AM, Mark Korver wrote:
I think current S3 pricing for 50TB at 0.125/GB comes to about
6250/month. That is for Designed for 99.9% Durability. I
can't even count that many 9s. I know you can buy HDs for about
there is a lot analysis of EBS IO performance out there. like
http://orion.heroku.com/past/2009/7/29/io_performance_on_ebs/
but I think, my earlier question, about the purpose of this mapserver
system needs to be addressed to be able to go further.
for example, if you are reading a lot of
Dont know if this relates to your problem, but its easy to check. We
normally use the GD/JPEG driver when doing image data. My
understanding is that AGG is more useful for outputting vectors to
PNG.
Also, this will slow it down but in the raster layer
PROCESSING RESAMPLE=BICUBIC rather than
I think your halos are blacks that are not pure black. Not 0/0/0,
just close. This is a compression artifact. I think the same thing
also happens with jp2 CCMs.
The NAIP CCMs/or County Compressed Mosaics were never meant to be
mosaicked together like this. They are a deliverable at the county
Many things come into play to make the response time of MapServer long
or short. But at a basic level the time required to get just the part
of the image required in the source file(s) and the CPU time required
to decode that part if compressed, are major factors in performance.
ECW is wavelet,
If all this is correct, I wondered if, for performance issues, it's better
to reproject vectorial drawings with MapServer or OpenLayers.
Depends on several things. If you just have small number of users on
the system, MapServer may well give you all the performance you need.
Plus its easy to
We all know that benchmarking is no easy task. Benchmarking when you
don't have the hardware in the first place makes things just a bit
more difficult.
I know this is not exactly the same thing. But cloud services, like
Amazon EC2, which give you access to a number of base configurations,
with
36 matches
Mail list logo
