@aaime I was checking solution for this one: https://osgeo-org.atlassian.net/browse/GEOS-9971 as we have an AWS environment where we have to give authorization to S3 through instance roles and are not allowed to create IAM keys.
It seems to work by changing ".setRequired(true)" to ".setRequired(false)", but it leaves *-sign to fields. Now I'm checking if * can be removed by leaving out whole ".setRequired".
What do you think, how should we proceed with this? Is it something that would go through as a pull request?
Ok. Do you have some example of a good way to make that configurable?
It seems I didn't get rid of that * - it comes from src\extension\gwc-s3\src\main\resources\GeoServerApplication.properties.
Anyways, authorization seems to work as it should with following cases. I created two buckets, bucket1 allowing access with IAM role for my instance and on another account bucket2 allowing access with IAM Keys
- No IAM-keys & bucket1 -> OK. Uses instance role
- IAM Keys & Bucket2 > OK. Uses IAM Keys
- Wrong IAM Keys & Bucket2 -> AWS API returns signature failure
- No IAM-keys & Bucket2 -> AWS API returns Forbidden
If I have understood right how AWS authentication chain works, client checks if following are available in this order:
- given IAM keys, empty in this case.
- credentials-file (~/.aws/credentials), in my case empty. As I had also ~/.aws/config empty, Geoserver managed to create a bucket to different region. So it seems, that buckets should be created before setting up BlobStore.
- Assumed role from the service, in my case ExecutionRole from ECS-service or InstanceRole for EC2-instance.
<filter-name>CorsFilter</filter-name>
<filter-class>org.apache.catalina.filters.CorsFilter</filter-class>
<init-param>
<param-name>cors.allowed.origins</param-name>
<param-value></param-value>
</init-param>
</filter>
<filter-mapping>
<filter-name>CorsFilter</filter-name>
<url-pattern>/</url-pattern>
</filter-mapping>
Hi guys. I'm running geoserver using tomcat9. our programmer develop a apps using leaflet with api and try to get the data from geoserver and an error cors. I amend the web.xml file at tomcat 9 as >> <filter>
<filter-name>CorsFilter</filter-name>
<filter-class>org.apache.catalina.filters.CorsFilter</filter-class>
<init-param>
<param-name>cors.allowed.origins</param-name>
<param-value></param-value>
</init-param>
</filter>
<filter-mapping>
<filter-name>CorsFilter</filter-name>
<url-pattern>/</url-pattern>
</filter-mapping>
can somebody help about this cors..after restart the tomact and geoserver..the error still same.. pls advise guys...thanks
Hello, on a fresh 2.18.2 geoserver instance publishing a layer fails with following error 500:
Handler dispatch failed; nested exception is java.lang.NoSuchMethodError: org.geoserver.geofence.services.dto.AccessInfo.getClipAreaWkt()Ljava/lang/String;Full error message here: https://hastebin.com/xilupocuri.apache Verbose logging did not show any useful in logs. Does one has any idea what is going wrong and why getClipAreaWkt() is missing?
on the working instance from yesterday I do have gs-geofence-server-2.18-20210225.224053-83.jar the installed version today is gs-geofence-2.18-20210517.135648-138.jar which is failing. I guess the numbers behind 2.18 are some timestamp which could indicate a different build?
@rappidGIS this is how I did some time ago. https://gist.github.com/t-book/e86fb4006e47d87acc64ad3b64c0b352 maybe my web.xml can help you
@t-book thank you sir..will check it out
Now I have Geoserver with gwc-s3-extension in AWS, connected to S3-bucket without credentials (see GEOS-9971) . It required a little bit of adjusting the platforms IAM-roles: EC2 (used for setting the configuration) requires InstanceRole with policy allowing some S3-operations. ECS Fargate (docker containers) required TaskExecutionRole with similar S3-policy. I have check what are minimal required S3-privileges, or has someone already sorted that out?
I think that kind of thinks would be helpful to be documented, but what's right place for them? Geoserver extension documentation or maybe a Github Gist about "running Geoserver in the cloud"?
Hi guys, i have a question about "reload band definition" in edit layer page on WEB UI.
What does it do? Because basically i have a layer associated to a custom store that retrive data from my BE.
Now, initially the store is empty (no image on it), then i upload it .... The point is, when i do "compute from data" (for the bounding box) it actually download the image from the store and read the image to get the BBOX.... But when i click on "reload band definition" it's too fast, there is no download, and put "empty" as band.
Differently if i create the layer when the store has image on it, the band are loaded right....
Why? What does "reload band definition" should do? Thanks
P.S. in order to create the layer without having an image on the store, my plugin if the store is empty create a 100 x 100 black image so geoserver doesn't go in nullPointer...
4 replies
Filter used to lookup user running as stated in the official tutorial.
4 replies