pgAgent Stumbling Blocks

In addition to doing scheduled backups from Postgres, we've also needed to load nightly dumps sent via scp to our database server's filesystem.  I finally got pgAgent working, but it wasn't without much consternation.  If you need to do this sort of thing on Windows, I'd recommend using Task Scheduler and a batch file if pgdump will work (though it likely won't if the file is coming from Oracle), or Task Scheduler with pgsql if it won't.  If you need to use pgAgent, here are some stumbling blocks that I found:

good support doc here: http://www.pgadmin.org/docs/1.8/pgagent-install.html

• add pgpass.conf to C:\Users\postgres\AppData\Roaming\postgresql
• should be localhost:5432:*:postgres:PASSWORD 
• (note, this is not under the default user, it's the "postgres" user on your system) 
• may also need the following in C:\Program Files (x86)\PostgreSQL\9.0\data\pg_hba.conf (but don't think so) 
• local all all trust
• host all all XXX.XXX.XXX.XXX/32 trust 
• reload configuration *running as admin* 
• install pgAgent via application/stackbuilder (*run as admin*) 
• make your steps ignore error, for some reason they error even if they complete successfully and stall subsequent steps

NOAA/NCDC CLIMAPS

National Climatic Data Center's Climate Maps of the United States database offers polygon data for many popular climate variables.  We would like to use this data for preliminary visualizations on bridge decay research -- the primary benefit to this data being its simplicity (it's already averaged by day/year over a 40 year period), national continuity, and file format (shapefile).

However we encountered two issues with the data.

Issue #1: it's based on a funky datum "GCS_NAD_1927_CGQ77", which addresses some issues in Quebec with NAD 1927.

We were able to side-step this issue by just re-defining the projection as something more common ... in our case we just went right to WGS 1984, since we wanted to view in Google Earth, and since a rough visualization was our immediate purpose.  I suppose NAD 1927 would have been fine, since we weren't doing anything with Quebec.  BTW, there are no readily available transformations of the  GCS_NAD_1927_CGQ77 datum.

Issue #2: the data is not continuous or even numerical

Values are expressed as "0 - 5 in", so we have no way of redefining what values are grouped in what colors, and since it's polygon data we also lose the ability to redefine spatial grouping based on new intervals.  That's not a big problem for this first rough visualization, but we will need to find more fine-grained data for subsequent analysis.  We'll probably end up taking point data from other NCDC sources and doing our own interpolation and temporal reduction.

ArcReader 9/ArcGIS 10 Conflict

I've seen a few folks who've had trouble installing ArcGIS 10 because they've been unable to uninstall ArcReader (usually version 9.x).

The best way I've found to do this is to download the ArcReader installer from here.  Run the installer and try to uninstall.

If you're unable to do that (sometimes you'll need to force quit and restart) you may need to run an install, but uncheck the ArcReader feature.  This gets rid of some hanging installer memory.

Then try to run a repair (restart if necessary) and start the downloaded ArcReader installer ... select repair when prompted.

After you've repaired you should be able to uninstall correctly.

If that doesn't work, you're stuck deleting all ArcReader files and ArcReader (and ArcGIS/Esri) references in the registry using regedit.

Delaware GIS 2012

The Delaware 2012 GIS conference was held yesterday.  It was a great opportunity to meet and learn from our colleagues in Delaware and the region.  I presented on Geospatial Data in Cloud and High Performance Environments, which you can access here: http://udel.academia.edu/BenjaminMearns/Talks/81729/Geospatial_Data_in_Cloud_and_High_Performance_Computing_Environments

Lewes Wind Turbine

I took a trip to the UD Lewes Wind Turbine on Thurdsday ... I posted a few paragraphs and a VIDEO to Doug White's Blog

ENVI/IDL and ArcGIS/Python integration

I occasionally answer questions on ENVI/IDL and ArcGIS/Python integration.

ENVI functions are available as "tools" through the ArcGIS GUI toolbox:

"With ENVI 4.8, ENVI’s scientifically proven image analysis capabilities will be available as discrete tools in the ArcGIS toolbox and accessible directly from ArcGIS desktop and server environments. Upon installation of ENVI products, a selection of powerful image processing and analysis tools will be available within ArcGIS."

A list of those tools is here (in the context of licensing, but just ignore that): http://www.ittvis.com/ProductsServices/ENVI/ToolsLicensing.aspx 




IDL can also be accessed for extension of ENVI/IDL functionality:

"developing customized ENVI tools can be done using simple IDL scripting and wrappers that facilitate the integration of ENVI tools with the ArcGIS platform."

Here's a whitepaper: http://www.ittvis.com/portals/0/whitepapers/ENVI_Tools_ArcGIS_Server_WP_Tutorial.pdf

Configuring svn on Apache

I followed this page to install/configure svn on my server ... it's written for Windows-based Apache install, so I've summarized and transcribed the following platform-anostic (and maybe overly general) steps:

1. make sure dav_module, dav_svn_module, and authz_svn_module are installed on Apache (assuming you're using webdav to do svn)
2. create an (arbitrarily named/located) svn config file (maybe called subversion.conf) and include in httpd.conf
3. create an (arbitrarily named/located) auth file (maybe called svn-auth-file) with the htpasswd binary that comes with Apache (should be under apache/bin), using -cm flag for first user and -m for each additional
4. create an (arbitrarily named/located) acl file (maybe called svn-acl) and assign permissions to users created in the previous step to groups and "projects" (directories) following this format:

# specify groups here
#
[groups]
team1 = ross, rachel

#
# team1 group has a read/write access to project1 repository
# all subdirectories
# all others have read access only 
#
[project1:/]
@team1 = rw
* = r

#
# project2 repository, only harry and sally have read-write access to project2
#
[project2:/]
harry = rw
sally = rw
* = r

4. inside the subversion.conf create a virtual directory using the following format, including the acl file created in the previous step

<location project1="">
  DAV svn
  SVNPath C:/Repositories/project1

  AuthType Basic
  AuthName "Subversion Project1 repository"
  AuthUserFile c:/etc/svn-auth-file

  Require valid-user

  AuthzSVNAccessFile c:/etc/svn-acl 
 </location>