-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Failed import causes unrecoverable error #354
Comments
Tried on a second fresh instance, downloading again from haitidata.org and got this:
|
Ok, new approach: #1. Let's install a vanilla GeoNode but not do anything with it.
After setting up the environment, let's re-try the data upload:
Same problem with vanilla geoserver-2.1.4: |
Full geoserver log is available at: |
I reported this in the GeoServer tracker, this issue is a blocker for the 1.2 release. |
This is unlikely to get fixed in GeoServer any time soon. Postgres has a built-in limit on the size of table names. You can change it at compile time (of Postgres) but it defaults to 64. GeoServer doesn't handle well the error when you ask it to create a table with an invalid name, but it probably makes sense to munge on the GeoNode side rather than try and get Postgres to allow unlimited-length table names. |
Thanks @dwins for your comment. I'd say the issue is not getting Postgres to allow unlimited-length table names, but rather that in this case, the result of a failed action in geoserver, makes subsequent actions fail. My post to the geoserver list did not get any comments in the last two months and my attempts to use different (non-community) channels to direct attention from geoserver developers to that report did not yield any results either. Examples like this of places where geonode usage hits obscure geoserver behavior, and our lack of ability to get the geoserver community interested in solving it is a liability to the geonode project and perhaps a good reason to consider multiple backends for the WMS / WCS / WFS, like we did with CSW. I will take your advice and implement the workaround you suggest on the python side. |
The actual error is:
If importlayers is cancelled and resumed, then the one of the failed layers can be properly imported:
A second attempt to import causes the error to be repeated:
An attempt to import only the offending one causes an error:
A second attempt causes a different error:
Restarting postgres does not fix the issue:
This is how the GeoServer log looks like:
The warning says there is code leaving feature readers/iterators open leaking statements and connections, let's look above in the logs to try to find out what the first exception looked like:
Since there is code leaking connections, let's try restarting tomcat to see if layers can be uploaded again:
It made me wait one minute, and then it failed:
Datasets used can be downloaded from http://haitidata.org
The datasets used can be downloaded from http://haitidata.org
The text was updated successfully, but these errors were encountered: