[mapguide-users] SHP Files Strike Again

[Dave Wilson]

Restarting the server should clear the cache yes. This sounds similar to
the issue raised by Jon R on the file I/O. I think we have recently just
fixed a bug where the connections weren't getting closed properly which
might explain the locking. If you are seeing a similar issue I suspect
it's the same bug. I can see this applying to both SDF and SHP files.

 

As far as managed connections to SHP you can't really point to multiple
SHP files in a single resource. You can LOAD multiple SHP files or
multiple SDF files, but it creates one FeatureSource per file. Unmanaged
connections is the only supported way to point a FeatureSource to
multiple SHP files similar to having multiple tables in a database or
multiple Feature Classes in an SDF. You might be able to modify the XML
for a managed resource to add multiple SHP files, but that was not how
the system was intended to work if it does at all - I've never tried it.
There is a difference in behavior because the system needs to enumerate
and perform a DescribeSchema on all the classes. In the case of multiple
files in the folder this could be a big task if you have many SHP files
in the folder. I'm not sure what the typical dataset is, but I suspect
it's more than 5 or 10 files and all their supporting files. The
DescribeSchema on a single file is quick.

 

It appears there are likely some performance issues with the folder
based functionality. The locking issue is a separate bug altogether
regardless of the number of files or type of connection involved as the
lock pertains to the specific file that has been opened for reading.
Rebuilding MGOS from current code likely resolves that problem if that's
an option.

 

Regards,

Dave

 

 

From: mapguide-users-bounces at lists.osgeo.org
[mailto:mapguide-users-bounces at lists.osgeo.org] On Behalf Of Chris
Gountanis
Sent: Monday, June 04, 2007 2:38 PM
To: 'MapGuide Users Mail List'
Subject: RE: [mapguide-users] SHP Files Strike Again

 

I guess my question is, where can you clear the pool cache when pooling
is enabled. Does a server restart handle it for you? I think it is
causing conflicts with SHP files when they are updated or replaced.

 

 

Dave,

I have moved to a single file per folder approach. Keep in mind the only
difference between managed and unmanaged connections is where the folder
located. Whether you have a single SHP or multiple files per folder that
is the problem.

 

 

Chris

  _____  

From: mapguide-users-bounces at lists.osgeo.org
[mailto:mapguide-users-bounces at lists.osgeo.org] On Behalf Of Dave Wilson
Sent: Monday, June 04, 2007 1:57 PM
To: MapGuide Users Mail List
Subject: RE: [mapguide-users] SHP Files Strike Again

 

Chris are you still trying to connect to a folder for SHP? It may be
that that is part of the issue. It might be worth falling back to a
resource per SHP file just like you have to do with SDF. 

 

Regards,

Dave

 

From: mapguide-users-bounces at lists.osgeo.org
[mailto:mapguide-users-bounces at lists.osgeo.org] On Behalf Of Chris
Gountanis
Sent: Monday, June 04, 2007 12:14 PM
To: 'MapGuide Users Mail List'
Subject: [mapguide-users] SHP Files Strike Again

 

I was playing around with my unmanaged SHP files and now the server just
crashes on zoom. I created a separate unmanaged feature source to see
if it was faster zoom wise and that seemed to cause the issue. I also
edited the layer to use new feature source with single SHP file and the
preview in Studio works fine. When bring the whole site up with AJAX
viewer the issue shows its face. Is there a reason why or a way to clear
cache? I had turned on SHP cache for pooling tests maybe that caused an
issue? I am not sure what the server is doing but it using 100%
resources then crashes on zooms now.

 

Chris

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/mapguide-users/attachments/20070604/64b4fce3/attachment.html