Re: [Geoserver-users] Many ImageMosaics based on a single db table
Hi Jakub Le 02/03/2020 à 14:15, Jakub Sawicki a écrit : I would like to keep a single table, but configure different mosaics (stores or layers) based on the column product_name. We have all our imagemosaic in a single table, and we have simply created one view in PostgreSQL for each specific imagemosaic. This way, you present the view to geoserver as if it were your base table, and you can very easily have on layer per view. Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Geowebcache rest/truncate guarantee of succcess
Le 24/09/2019 à 15:19, Nuno Oliveira a écrit : If you manage to reproduce the issue, ideally whit GeoServer release data, then feel free to open a JIRA ticket. I surelly will. But it happened only once (or I did not realize it the other times), and it happened on an unmanaged server so I'm not sure I will easily replicate this. Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Geowebcache rest/truncate guarantee of succcess
Le 24/09/2019 à 12:50, Nuno Oliveira a écrit : That say, when you get back a 200 and the cache on disk was not deleted, it usually means that one of the provided truncate parameters didn't match the cache (e.g. style, image format, etc ...). I already got a truncate request that did not delete the cache on disk, even though the parameters were the same as usual (the request is scripted, and it worked before and after the failure). I could not swear that I had gotten a 200 back but I'm 100% sure that the parameters were correct. Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Clustering query results
Hi Walter Le 25/03/2019 à 01:18, wambac...@posteo.de a écrit : select st_centroid(unnest(st_clusterwithin(geom,0.00045))) geom from assoc3_house This query is used in a sql-layer (WFS) and gives reasonable results. :) BUT: the performance is very, very poor. Tablesize is about 460.000 nodes and there is a spatial index on geom. *Runtime 3-4 minutes.* In PSQL the same query (of course with the corrisponding bbox) is running in 2-3 seconds giving the same results. Have you rewritten the query, or have you copy/pasted it from GeoServer's log ? Geoserver wraps the query inside a subquery and then applies the bounding box's where clause on the subquery. Maybe this is what causes the different behaviour. Note that if you do : select st_centroid(unnest(st_clusterwithin(geom,0.00045))) geom from assoc3_house where geom && bbox you will filter on the base geometry and therefore use your spatial index on the "geom" column. But if you do : SELECT * FROM (select st_centroid(unnest(st_clusterwithin(geom,0.00045))) geom from assoc3_house) AS vtable WHERE geom && bbox you will filter on the centroid geometry which is the subqueries output and therefore NOT the index on the geom column. Both queries wont give exactly the same results, for clusters that don't entirely lie within your bbox for instance. Cheers ! -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] GeoServer 2.16.x - JDK11 - JAI Image IO
Hi Florian Le 11/02/2019 à 17:40, geomatick a écrit : I installed GeoServer 2.15-RC in Tomcat 9 in Windows 64B, I know this GeoServer version is in development. Are you using a 64bit version of Java ? Because there is no native 64bit version of native JAI or JAI_ImageIO on Windows. Anyway, depending on your workload, you might still find it better to use 64bit Java without JAI (that's the case for us because we have to cope with very memory hungry WMS requests so we need very big heap sizes). Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid "sorry too many clients"
Hi Andrea Le 07/02/2019 à 10:09, Andrea Aime a écrit : Hum... the schema in the datastore configuration should affect only the feature types served by the datastore itself. Are you expecting it to affect the mosaics somehow? No, sorry, I got lost in the process. Now I believe I understand. For some reason I thought that the JNDI datastore would be used by the imagemosaic. But the imagemosaic directely references the JNDI connection, not the JNDI datastore. Makes sense. Did you setup any connection between the datastore config and the mosaic? No, that's what I thought was happening when I set up the jndiReferenceName in the datastore.properties file to be the same as the jndiReferenceName in the JNDI datastore configuration. Now I understand that both refernce the jndiReferenceName defined in the context.xml of my tomcat server. Totally logical, but as you can see I'm not very knowledgeable in that matter. I'm afraid not, while the generated output of the training material is under a creative commons license, the source code of it is in a private repo. But it's a good suggestion, I've opened a ticket to remind us a fix the next time we work on the material. OK, thanks for your help Andrea. Cheers ! -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid "sorry too many clients"
Le 05/02/2019 à 12:14, Andrea Aime a écrit : for that case you need to "step up" your installation and have everything use the same JNDI connection pool (same connection pool for all stores), see also: https://docs.geoserver.org/stable/en/user/tutorials/tomcat-jndi/tomcat-jndi.html JNDI works like a charm ! Thanks ! Just a note, I don't know whether this is a bug or something missing in the documentation, or maybe even something I have overlooked : the "schema" parameter in the Datastore Configuration page does not seem to have any effect. I've set it to the name of my postgresql's database schema, but the log was flooded with geotools.jdbc trying to read the primary keys of all the tables in my database (and not in my schema alone). I had not checked "expose primary keys". More problematic, geoserver that the relation containing my mosaic index did not exist. In the datastore.properties of each mosaic (and of my imagepyramid) I added "schema=myschema", and everything worked perfectly afterwards. Is the the expected behaviour ? Should I propose a quick patch in the jndi configuration doc ? Also for my datastore.properties file, I had followed the guide from the "geoserver training" section of your website : https://geoserver.geo-solutions.it/edu/en/multidim/imagemosaic/mosaic_datastore.html#using-a-jndi-connection-pool which did not mention this "schema" parameter, so mabe I could add this somewhere ? Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid "sorry too many clients"
Hi Andrea Le 05/02/2019 à 12:14, Andrea Aime a écrit : for that case you need to "step up" your installation and have everything use the same JNDI connection pool (same connection pool for all stores), see also: https://docs.geoserver.org/stable/en/user/tutorials/tomcat-jndi/tomcat-jndi.html That looks like a good option. For the time being I was thinking about configuring a pgBouncer service (i.e. connection pooling middleware), which I guess would achieve the same result. But I'll look into JNDI first if that can spare me an additional software to maintain. Thanks for the suggestion ! Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] ImagePyramid "sorry too many clients"
Hi list, I have an imagepyramid made of of 16 imagemosaic layers. The imagemosaic indexes are stored in a postgis database. I've had connection exhaustion problems (i.e. "sorry, too many clients" from the PostgreSQL backend). My guess is that even though the imagepyramid has its own datastore.properties file, each imagemosaic layer uses the "max\ connections" parameter from its own datastore.properties file. So for 16 imagemosaic layers with the default max connections setting (10), we could go as high as 160 idle open connections. Is this assumption correct ? If I am right, then the only way to keep these connections at a reasonable level would be to set the max connections per imagemosaic to something like 4 or 5. Thanks for your advices on this ! Cheers ! -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] one layer two database queries
Le 28/01/2019 à 09:26, Andrea Aime a écrit : There was a mention of this topic in "GeoServer on steroids, slide 25": https://www.slideshare.net/geosolutions/geoserver-on-steroids-foss4g-2015 And also in GeoSolutions performance optimization training material: https://geoserver.geo-solutions.it/edu/en/enterprise/styles.html#the-concept-of-featuretypestyle Both are must reads for anybody planning a deployment of GeoServer in a production environment ! Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] one layer two database queries
Hi Peter Le 25/01/2019 à 20:15, Peter Smythe a écrit : Perhaps you can update the documentation at https://docs.geoserver.org/latest/en/user/styling/sld/extensions/z-order/example.html with the lessons that you have learnt, especially the memory impact and the multiple queries. Then we can all benefit from your experience. If you are not familiar with contributing on GitHub, send me the changes in a Word document and I will gladly help you. Not very familiar indeed, but enough to contribute I guess. When this project is over I'll have more time to document things and give afterthougts. I'm not sure that the impact of multiple FeatureTypeStyles on memory would fit in the z-order documentation though. This is a more fundamental concept of GeoServer. Maybe this problem of FTS and memory is not even a subject for the documentation, but maybe a good subject for a blog post for instance. I remember reading a documentation page on the subject, but can't get my hand on it anymore. There was a drawing of multiple drawing canvas being merged together. Maybe this was in relation with composition, but even there I haven't found anything... Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] one layer two database queries
Le 25/01/2019 à 14:12, Andrea Aime a écrit : In if you have casing, a minimum of two groups is needed. See documentation here, it shows CSS for brevity but also links to a SLD: https://docs.geoserver.org/latest/en/user/styling/sld/extensions/z-order/example.html Yes, actually that's the example I followed for my own style .But I did not understand the memory impact of having many FeatureTypeStyles and I found it very convenient, so I went with on FTS per road class... A lot of SLD rewriting in perspective ! Thanks a lot for these great tips Andrea ! Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] X-Frame Options
Hi Wolfgang Le 25/01/2019 à 09:09, Landratsamt AS (SG 11.2 - GIS) a écrit : another viewer wants to display a wms from me. Unfortunately, the factual data does not work. Mozilla Firebug says: "Load denied by X-Frame Options" An adjustment in web.xml does not work: I've had this problem some weeks ago, and the easy solution was to use an environment variable. The env variable should be named "geoserver.xframe.shouldSetPolicy" and its value should be "false" (both without quotes of course). You can find the discussion here : https://sourceforge.net/p/geoserver/mailman/message/36468205/ Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] one layer two database queries
Le 25/01/2019 à 10:34, Andrea Aime a écrit : Oh dear, now I see it... and it's truly horrible! :-D I thought the various FeatureTypeStyle had a different FeatureTypeName, but it's just a different Name instead. FeatureTypeStyles are very heavy objects rendering wise, you should not use them as mere classifiers. Every time you create a FeatureTypeStyle, a new z level in the map is created. So creating a new z level for each FeatureTypeStyle is pretty much what was intended. This means that even in the best of conditions, with that style GeoServer would be allocating tens of drawing surfaces, with an insane memory usage going with it (I think I see 42 FeatureTypeStyles, for a 4x4 metatile , 1024x1024, 4 bytes per pixel, that would be 168MB of RAM (you don't get that high, see below, the stacks of rendering surfaces are actually split in two). Now, that's indeed very bad. I do need to rewrite things. So why is it doing two queries anyways? Because you have two different names for the sortGroup, which makes the various feature type styles un-mergeable. Oh OK, now I see it ! Nice pick Andrea ! Do yourself a favor and, if possible, use a single FeatureTypeStyle instead ;-) OK, I understand the problem now, very clear. The problem is that I do need to order all these features the way I did (i.e., for instance, 'water' on top of 'wood'). So, from what you say, the "right" way to do this would be to use a single FeatureTypeStyle for the first sortByGroup, and use a zOrder and add the appropriate column in my datasource ? For the second sortByGroup, I need to have many things ordered (i.e. road casing and roads going over/under buildings. So in this case, I believe I do need one FeatureTypeStyle per road type and a sortbygroup, right ? Thanks a lot for your help Andrea ! Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] one layer two database queries
Le 24/01/2019 à 14:11, Andrea Aime a écrit : Wow, it's big... and the feature type names are not a match for the table names you're using, seems like you're using SQL views? Yes, SQL view, but basically only selecting all the columns (and doing simple CASE ... WHEN statement to do things that would be more difficult in SLD) and filtering on a boolean field. And while the target table is the same, the sql view definition is not. What do you mean ? The SQL view definition is the same, since it's only one single layer. What is different though is what is added by geoserver, i.e. whats is after the "AS vtable". But everything inside the subquery (i.e. the view definition) is the same. More in general, there are no facilities to figure out that two feature type styles are based on the same base table and eventually reorganize the queries accordingly, the renderer only sees the feature types, has no idea what's behind them, and queries them one by one. Well, that is, unless you mean that a specific user visible layer, with one name, is causing two different queries? But I don't see that from your query logs. Yes, that's exactly what I mean. I query a single layer (base on the SQL view that you can see inside the "vtable" subquery) using the URL that I previously posted. Geoserver then issues two different queries, the first time with a part of the feature type styles, and the second type with the rest of it. Actually the style is not that big (very subjective, of course), it's a rendering based on a very simplified version of openstreetmap's "planet_osm_polygon" table, with a lot less features than on OSM's own style. Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] one layer two database queries
Le 24/01/2019 à 13:09, Andrea Aime a écrit : On Thu, Jan 24, 2019 at 12:37 PM Arnaud L. <mailto:arnaud.lis...@codata.eu>> wrote: The SLD (actually it's YSLD, hope it's fine) For me, it isn't, I really dislike that language. No problem, here is the same style in SLD : https://pastebin.com/Nsief1vf Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] one layer two database queries
Hi Andrea Le 24/01/2019 à 12:22, Andrea Aime a écrit : can you share the exact request you're making and the full SLD involved? Sure ! The query is : http://geoserver:/geoserver/wms?SERVICE=WMS=1.3.0=GetMap=image%2Fpng=false==false=EPSG:32631=osm:web_polygons=1200=1200=130683,482,132207,4824857 The SLD (actually it's YSLD, hope it's fine) is here : https://pastebin.com/TQwzu4Pb Also, here are the two queries (not sure they are in this order, I think they are reversed) : https://pastebin.com/CryUBDbu I've set a one day expiration on the two pastes. Cheers ! -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] one layer two database queries
Hi list I see in the debug log that when I query a specific layer via WMS, geoserver queries my postgis table twice. The bbox for both queries is slightly different because the stroke width for both queries is different. It seems like this is unnecesssary overhead to me. What is causing this "double query", and how can I fix it ? Thanks ! Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Resolving issues with duplicated labels in meta-tiling
Le 21/01/2019 à 10:37, Eugene Podshivalov a écrit : Sorry, yes, your are right, it does work but with point (centroid) labeling you lose the possibility to render the label relative to the polygon size. You will need to define the arae-zoom ranges manually in the style, won't you? Yes exactly, that's what I do. I have an area field and I draw labels depending on the zoom/area combination. You lose a lot of flexibility, but that's the only solution that worked 100% of the time for me. Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Resolving issues with duplicated labels in meta-tiling
Hi Eugene Le 21/01/2019 à 09:57, Eugene Podshivalov a écrit : That solution works only for un-tiled requests. Maybe I haven't understood you requirements then, because I use this 'trick' on some tiled wms request and it works great for me. It guarantees a single label location, however you query your layer. Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Resolving issues with duplicated labels in meta-tiling
Hi Eugene Le 20/01/2019 à 18:52, Eugene Podshivalov a écrit : Meta-tiling resolves most of the issues with duplicated labels but they are still happening on geometries which cover several meta-tiles. You can see that illustrated on the diagram below. Would it be possible to resolve it by putting a label only to that meta-tile which covers the largest portion of the geometry? This is the kind of problem that could be fixed by placing the label on the centroid of the polygon instead of using automatic placement (which would position the label on the centroid of the *visible* portion of the polygon). You can see a solution by Andrea here for instance : https://gis.stackexchange.com/questions/265836/label-polygon-at-central-position-dynamically Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Dynamic add for geotiff layer
Hi Marten Le 16/01/2019 à 14:13, Mårten Swärd a écrit : Im currently working on a Web application that is supposed to allow a user to upload a geotiff to the system and then display it on a map. How do I programatically add a geotiff to Geoserver so that I can get an accessible layer thru WMS? I would create an imagemosaic layer with a postgis index. This way, all you'd have to do is save the geotiff on your server and add an entry in your database index table to point to this geotiff. Quite straightforward I believe. Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] WMS fails in SVG but works in other formats
Hi Andrea, thanks for your answer ! Le 09/01/2019 à 09:43, Andrea Aime a écrit : This is like saying "If I go to work by car or bike everything works, but using the train, something breaks, any idea what might be causing this?". The code writing out the format is very different in these three formats, PNG writes using Java2D, PDF using iText and SVG using Batik. I gave a stack trace, so I feel I did a little bit more than that... Making a guess, maybe the style is pointing to an image that cannot be located, and SVG will throw an NPE instead of silently skipping it? The png files are there (I checked them all, they were not many) and are rendered fine with the other encoders. Or, if this is the actual code (lines number might not match, but it seems they do) it could be that for some reason Batik is not finding a PNG encoder (the writer variable), but I would not know why or how: https://github.com/apache/batik/blob/trunk/batik-svggen/src/main/java/org/apache/batik/svggen/ImageHandlerBase64Encoder.java#L152 Maybe, but this is definitively beyond my knowledge. This would need someone running a debugger inside the Batik encoding (and possibly initialization of that registry) and figure out what happened. A first step would be to open a bug report on Jira with reproducible steps (including data and style needed). Also, make sure you've tried the generation on the latest stable version (2.14.1) as we have been upgrading Batik a few times recently. Yes, I'll do this ASAP. I'll first upgrade to 2.14.1 to make sure this is not something that has already been fixed. Thanks for your help ! Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] WMS fails in SVG but works in other formats
Hi Brad, and thanks for your answer Le 09/01/2019 à 09:43, br...@frogmouth.net a écrit : Looking at the code, it's a rendering error, but the stack trace isn't really that informative. Do you have a specific case that reproduces this? Any chance that you can debug it to find out the specific rendering exception that occurs (probably in Batik, maybe in GeoTools)? I can reproduce this by trying to render any layer that has a visible png symbolizer. I don't know much (almost nothing, actually) about java coding/debugging, so it's not an easy task for me. I'll be glad to dig into this when I have some spare time though (but this is not the cas at the moment unfortunatelly). Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] WMS fails in SVG but works in other formats
Hi Ian, thanks for your answer Le 09/01/2019 à 09:42, Ian Turton a écrit : Is this the case for all your layers or just one? All layers with external png images as symbolizers apparently. Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] WMS fails in SVG but works in other formats
Hi list, and happy new year to the geoserver users and developers ! I have a wms request that works fine when the output format is set to either PDF (application/pdf) or PNG (image/png), but fails when trying to set the output format to SVG (image/svg). Any idea of what might be causing this ? Geoserver version is 2.14.0. Stack trace is here : https://pastebin.com/EHq0Lmau Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] WMS : rasterized layer in pdf output fprmat
Hi Andrea Le 15/12/2018 à 19:48, Andrea Aime a écrit : GeoServer is just calling an iText Graphics2D object as if it was doing a normal map paint, there is no "vector" or "raster" mode, whatever happens inside that iText object, happens, GeoServer just tries hard to pass as many vector primitives as possible to it. Of course, if a fill is made of raster icons, it cannot magically become vector, besides that, it may be that some vector primitives get rasterized by iText, e.g., I see a dashed line, if you used "stroke-dasharray" GeoServer is calling a vector primitive on the Graphics2D, but it may be iText is building a raster representation of it. I've been thinking about it, and I think it is much more simple than that : I use composition between layers (destination-in), and this of course can only work on rasters. So I believe this is where the rasterization happens. Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] labelObstacle and closed linestring
Le 14/12/2018 à 11:48, Andrea Aime a écrit : I don't see any new pull request though? OK I forgot the last step, "create pull request". Quite a newbie in this matter... Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] labelObstacle and closed linestring
Le 14/12/2018 à 11:48, Andrea Aime a écrit : On Fri, Dec 14, 2018 at 11:35 AM Arnaud L. <mailto:arnaud.lis...@codata.eu>> wrote: Le 14/12/2018 à 11:11, Andrea Aime a écrit : > Easy to miss indeed. Do you want to contribute a documentation fix, > moving that warning higher up in the page? Done. I moved the warning below the description and before the code. I don't see any new pull request though? OK, not familiar with github at all, I'll go over the process once again. -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] wms application/pdf dimensions of document
Hi List Trying to make simple PDFs using WMS GetMap requests. It works great, but I expected the document size to follow the format_option:dpi=XXX parameter. It is not the case, and in PDFMapResponse.java, I see : // step 1: creation of a document-object // width of document-object is width*72 inches // height of document-object is height*72 inches Why is this hardcoded, and where ? Wouldn't it be a good idea to use the dpi given in format_options to build the document ? If I understand the iText doc correctly, one could set the document and template width and height to width*72/dpi and height*72/dpi, and scale the Graphic2D as well to 72/dpi. This way, the pdf would still be the same but it's dimension would be coherent with the chosen DPI. Before you ask : I'd gladly propose a change in github, but I have almost zero knowledge of Java so I don't feel like I could handle this in a reasonable time... Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] labelObstacle and closed linestring
Le 14/12/2018 à 11:11, Andrea Aime a écrit : Easy to miss indeed. Do you want to contribute a documentation fix, moving that warning higher up in the page? Done. I moved the warning below the description and before the code. Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] labelObstacle and closed linestring
Le 14/12/2018 à 10:37, Andrea Aime a écrit : On Fri, Dec 14, 2018 at 10:18 AM Arnaud L. <mailto:arnaud.lis...@codata.eu>> wrote: My two cents is that labelobstacles are computed on bounding boxes, and therefore will never be drawn inside any kind of closed linestring. Am I correct ? Yes, exactly as documented, did you scroll to the bottom of the page about it? Obviously not, I missed the most important part... Sorry for the inconvenience. Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] labelObstacle and closed linestring
Hi List I'm trying to prevent street labels do be drawn on top of some polygon boundaries (they can be drawn inside, but not cross the boundary). I understand that using labelobstacle on a polygon feature prevents all labels crossing AND inside polygons, this is the obvious intended behavior. So I created a new layer with only the boundaries of the polygons as linestrings and only added the labelobstacle parameter on this layer, but to my surprise no label is drawn inside these linestrings either. Is this the expected behavior ? My two cents is that labelobstacles are computed on bounding boxes, and therefore will never be drawn inside any kind of closed linestring. Am I correct ? Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] WMS : rasterized layer in pdf output fprmat
Le 10/12/2018 à 12:16, Andrea Aime a écrit : On Mon, Dec 10, 2018 at 11:58 AM Arnaud L. <mailto:arnaud.lis...@codata.eu>> wrote: Here is a sample PDF, the white polygons are the rasterized ones : https://transfer.sh/ep3Ab/test.pdf The link does not work for me, it gets stuck in "connecting" status and never returns. I've made a quick check against a layer group we have on a demo server, there are multiple layers and they all seem to be vector (well, besides the background raster one of course ;-) ): http://cloudsdi.geo-solutions.it/geoserver/daraa/wms?SERVICE=WMS=1.1.1=GetMap=application%2Fpdf=true=daraa%3ADaraa-OSM=application%2Fvnd.ogc.se_inimage=EPSG%3A4326==768=648=36.07240676879883%2C32.59446144104004%2C36.13832473754883%2C32.65007972717285 Yes, sorry for the link, transfer.sh appears to be down. I attached the file since it is really lightweight. I have a guess though : in the second layer, some polygons have a raster graphic symbolizer (external png graphic). Could this trigger a layer rasterization ? Cheers -- Arnaud test.pdf Description: Adobe PDF document ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] SVG problem
Hi Andrea Le 26/11/2018 à 10:22, Andrea Aime a écrit : interesting finding, maybe you or Olyster could open a ticket with the sample SVG and an explanation of why it does not work? Where should I open this ticket precisely ? Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] SVG problem
Le 23/11/2018 à 21:58, Olyster a écrit : I tried every combination of parameters in sld but the svg is showed with no height. I mean it looks flatten. The svg opens fine in Inkscape. You should add width and height properties to your svg tag. This is undocumented and I don't know if it really is a requirement or not, but I had no success with SVGs that did not have these tags set. So in your case something like : http://purl.org/dc/elements/1.1/; xmlns:cc="http://creativecommons.org/ns#; xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#; xmlns:svg="http://www.w3.org/2000/svg; xmlns="http://www.w3.org/2000/svg; id="svg3761" version="1.1" viewBox="-17 0 35 18" width="35px" height="35px"> I've not found the correct settings neither in Inkscape nor in Illustrator to have them set these properties properly upon saving, so I do it manually. Inkscape does add width and heigt, but it also sets x and y which geoserver does not seem to like. Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] X-Frame-Options
Le 16/11/2018 à 09:13, Andrea Aime a écrit : geoserver.xframe.shouldSetPolicy false OK thanks Andrea. But then why not init-param in the xFrameOptionsFilter ? Sorry, I'm not very familiar with java environments... BTW, I used the env variable and it works fine ! Cheers ! -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] X-Frame-Options
Hi mark Le 15/11/2018 à 20:16, mark a écrit : actually this is wrong, it looks to parse a boolean from geoserver.xframe.shouldSetPolicy https://github.com/geoserver/geoserver/blob/94d8bca73b5e1a7438117b1aa2fb519b6ae4e3a4/src/main/src/main/java/org/geoserver/filters/XFrameOptionsFilter.java#L33-L35 Indeed, that's the property that I've (tried to) set to false. alternatively you could remove the filter, tough setting as an environment variable will be easier when upgrading Yes, that's what I ended up with. Thanks Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] X-Frame-Options
Hi list ! As per the documentation here : https://docs.geoserver.org/latest/en/user/production/config.html#x-frame-options-policy I have tried to disable the X-Frame-Options in web.xml (using Apache Tomcat 9.0), but no success. Is this the correct syntax ? xFrameOptionsFilter org.geoserver.filters.XFrameOptionsFilter geoserver.xframe.shouldSetPolicy false Thanks a lot ! Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Make hole in line to draw label
Le 14/11/2018 à 10:27, Andrea Aime a écrit : Labels are drawn last, on top of everything, so there is no possibility to composite them at the moment. That's another change that might work, but it plays badly with priority support and conflict resolution (both require labels to be accumulated and handled at the end), so it's not a trivial change to implement (and it's basically how "holes" in lines would have to be implemented anyways) OK, thanks for clarifying this. On this subject, allowing some textsymbolizers to be taken out of the priority/conflict resolution pipeline would also solve some other problems (drawing geometries on top of labels being the most obvious one). i.e. having something like an x-textasgeometry vendor option to take a particular feature out of the labels pipeline and simply be drawn with no conflict resolution just like any other feature. Anyway, thanks for all your answers ! Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] Make hole in line to draw label
Le 14/11/2018 à 10:18, Andrea Aime a écrit : There is no facility to add gaps around labels, but if someone were to develop or sponsor it, it would be a welcomed addition. I'd be happy to help on this topic as well as the YSLD one (see, I got the message ;D), but unfortunately my Java knowledge is way too basic to even think about contributing. The halo is not working also because it has little contrast with the label itself, it's white vs light gray. If the label was darker, it would be readable no matter what the background is. To me the label is not that readable on the white background either. Yes I know, it's only an example I'm still fine tuning this particular style (you can see a very readable black label in the bottom left of the picture). What about compositing with labels ? How does one achieve this, when labels are supposedly drawn last ? Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] Make hole in line to draw label
Hi list ! I'd like to label a line with a text in follow-line positionning. My problem is that with this simple style the label is not very readable because it is "stroked" by the line. So I'd like to make a hole (a gap), in the line at the place where the label will be drawn. I tried with a halo but it does not work well in my case because I have different background colors, and the halo has a predetermined color. So on white backgrounds a white halo works, but on blue background it does not. See here for a concrete example : https://i.imgur.com/39PQDX2.png Could anyone suggest me another way to achive this ? Thanks a lot Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] SLD styling in production environment
Le 08/11/2018 à 17:34, Johannes Weskamm a écrit : Just want to add another option for a simple, what you see is what you get SLD-editor: Demo: https://terrestris.github.io/geostyler/ Code and documentation: https://github.com/terrestris/geostyler Introduction: https://www.terrestris.de/en/2018/08/06/das-projekt-geostyler/ News: https://www.terrestris.de/en/2018/09/12/neues-vom-geostyler/ This project is still a work in progress, but it think its already usable! Very nice project ! Are there by chance any plans for YSLD/CSS formats ? Cheers -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] SLD styling in production environment
Le 07/11/2018 à 16:27, Olyster a écrit : to me YSLD (extension) offers more readability. I also work with YSLD. It's main advantage is that it is a 1-1 replacement for SLD, but way more compact and readable (and writable, since I write all the styles manually in a text editor). You can convert YSLD to SLD and vice-versa via the REST interface. Very useful. The problem is that the documentation is a bit harder to search thant the SLD documentation. Google indexing of the YSLD doc is not very good, and many examples that used to exist in YSLD are now in CSS. -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid layer black when no granule at all
Le 31/10/2018 à 15:21, Andrea Aime a écrit : Hum.. I cannot exclude it (but I'm not sure without checking the code). Did you provide a TypeName in the indexer, to tell the mosaic which table to use? Wondering if doing so might convince it to actually use the table in question. No, indeed. Excellent suggestion, because the first table it might encounter with the correct schema (i.e. location and the_geom fields) is the index for a 2.54 level mosaic ! I added the TypeName, and lo!, levels are correct ! Thanks a lot Andrea ! Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid layer black when no granule at all
Le 31/10/2018 à 14:59, Arnaud L. a écrit : Why would the imagemosaic plugin guess that levels are 2.54 ? Is there something wrong in the way I configured the imagemosaic ? OK, I have a guess. In the FOLDER referenced by the index there are some GeoTIFFs whose level is 2.54. None of these GeoTIFFs are referenced in the index though. Could it be that the imagemosaic plugin does a scan on the entire folder, even though I have set UseExistingSchema=true ? Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid layer black when no granule at all
Le 31/10/2018 à 14:03, Andrea Aime a écrit : No, you are still supposed to let the mosaic configure itself. Never, ever, generate the mosaic property file manually, setup a indexer and a datastore config files and let the mosaic do its work, regardless of where the index is stored. OK Andrea, thanks for making this clear. I did as suggested : - copied only the datastore.properties in a folder - added an indexer.properties in the same folder, basic configuration taken from https://geoserver.geo-solutions.it/edu/en/multidim/imagemosaic/mosaic_indexer.html, i.e.: Name=georasters_17 Schema=*the_geom:Polygon,location:String PropertyCollectors= Caching=false AbsolutePath=true CanBeEmpty=true UseExistingSchema=true - created a new imagemosaic datastore pointing to this folder Everything went fine, mosaic "georasters_17.properties" and "sample_image.dat" created successfully. Now, comparing this properties file with the one I created manually, there is one major difference (and some minor ones like expandToRGB), Levels! In the automatically created file : Levels=2.54,2.54 In my file, where I computed the levels myself : Levels=0.15875,0.15875 The "correct" one, i.e. the one that renders correctly is mine. Gdalinfo on the underlying GeoTIFFs gives the correct pixelsize, i.e : Pixel Size = (0.1587500,-0.1587500) Why would the imagemosaic plugin guess that levels are 2.54 ? Is there something wrong in the way I configured the imagemosaic ? Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid layer black when no granule at all
Hi Andrea Le 31/10/2018 à 13:00, Andrea Aime a écrit : thanks for sharing! It would have been hard to guess that you had a image mosaic without sample image, those files are mandatory and manual generation of the mosaic setup is uncommon. Is it ? But manual generation is mandatory when one uses a PostGIS based mosaic index, isn't it ? Are most users relying on shapefiles ? They are very impractical when dealing with dynamic data though. Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid layer black when no granule at all
Le 30/10/2018 à 14:04, Arnaud L. a écrit : When there are granules in the bbox, even when they only partially cover the area, the region where there is no granule is transparent. But when there are really no granule to load at all, the resulting PNG is fully black, with no transparency. [...] So how can I have a fully transparent PNG image when there is no granule in the queried bbox ? All right, I found the solution ! For the record : In the log, the problem obviously appeared when no granule was found, and the problematic part was : "[org.geotools.gce.imagemosaic] - Creating constant image for area with no data" This was logged by createBlankResponse() in RasterLayerResponse. Trying to understand what happened in there, I saw that it created a blank image based on the default color model and default sample model. Then I realized I had no sample_image in the underlying mosaics, because they are managed by my own processes. So the solution was simply to generate a sample_image file which would have the same RGBA color model as my dataset. For that I created a dummy imagemosaic datastore with a single GeoTIFF from my dataset. GeoServer automatically generated the index file and the sample_image.dat file. I copied this image at the root of all the imagemosaics in my pyramid, and it instantly worked ! Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] ImagePyramid layer black when no granule at all
Hi all I have set up an ImagePyramid layer (thanks Daniele) that I query via WMS format=image/png transparent=true It uses the default raster style. When there are granules in the bbox, even when they only partially cover the area, the region where there is no granule is transparent. But when there are really no granule to load at all, the resulting PNG is fully black, with no transparency. When using gdalinfo on the "normal" PNG image I have : Band 1 Block=1281x1 Type=Byte, ColorInterp=Red Mask Flags: PER_DATASET ALPHA Band 2 Block=1281x1 Type=Byte, ColorInterp=Green Mask Flags: PER_DATASET ALPHA Band 3 Block=1281x1 Type=Byte, ColorInterp=Blue Mask Flags: PER_DATASET ALPHA Band 4 Block=1281x1 Type=Byte, ColorInterp=Alpha When using gdalinfo on the "black" PNG image I have : Band 1 Block=1281x1 Type=Byte, ColorInterp=Gray Mask Flags: PER_DATASET ALPHA Band 2 Block=1281x1 Type=Byte, ColorInterp=Alpha I cannot set outputtransparentcolor to #00, because I have black areas in my granules that must not be rendered as transparent. I have found this old thread that seem to be relevent in my case : http://osgeo-org.1560.x6.nabble.com/ImageMosaic-transparency-problem-td3795094.html but I don'ty think the solution would be suitable for me, because I have no control over the input granules' palette. So how can I have a fully transparent PNG image when there is no granule in the queried bbox ? Thanks a lot for your help ! Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] sortByGroup in YSLD error
Hi Brad, thanks for your answer Le 26/10/2018 à 10:27, br...@frogmouth.net a écrit : YSLD examples can be found here: https://docs.geoserver.org/latest/en/user/styling/ysld/cookbook/index.html Yes I have read these pages, but they seem incomplete. The x-sortBy and x-sortByGroup properties are not mentionned anywhere in this documentation. Furthermore, the "z-ordering across FeatureTypeStyle" used to have example styles in YSLD. I can't find them anymore. Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] sortByGroup in YSLD error
Le 26/10/2018 à 09:21, Arnaud L. a écrit : What's wrong with this style ? I think I found the problem, but I don't understand it : ALL the featurestyles in the layer must have the x-sortByGroup property set. Why is it so ? That's not what I understand from the documentation. In https://docs.geoserver.org/latest/en/user/styling/sld/extensions/z-order/syntax.html#z-ordering-across-layers, it is clearly stated that the rails layer has "two FeatureTypeStyle, one with a group, the other not". PS : how come I can't find any YSLD example anymore ? The YSLD reference does not mention the vendor options like x-sortBy anymore, is this format being abandonned ? I'm still worried about that. Why are the YSLD examples gone ? It looks like they were replaced by CSS. Did I make a bad choice with YSLD ? Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] sortByGroup in YSLD error
Hi all I have two layers that I query together via WMS, and I'd like to order features between them as in the "z-ordering across layers" example in the doc. I style the layers in YSLD. In my first style, i have a bunch of FeatureTypeStyles, and only the last one has a "x-sortBy" property. In my second style, all FeatureTypeStyles have a "x-sortBy" property and a "x-sortByGroup" property. The problem is that as soon as I add the "x-sortByGroup" property in the first layer's style, WMS rendering breaks with this stack trace : java.lang.NullPointerException at org.geotools.renderer.lite.ZGroupLayerFactory.ensureSortProperties(ZGroupLayerFactory.java:175) at org.geotools.renderer.lite.ZGroupLayerFactory.buildNewFeatureLayer(ZGroupLayerFactory.java:159) at org.geotools.renderer.lite.ZGroupLayerFactory.addToSplitLayers(ZGroupLayerFactory.java:144) at org.geotools.renderer.lite.ZGroupLayerFactory.arrangeOnZGroups(ZGroupLayerFactory.java:121) at org.geotools.renderer.lite.ZGroupLayerFactory.filter(ZGroupLayerFactory.java:85) at org.geotools.renderer.lite.StreamingRenderer.paint(StreamingRenderer.java:803) at org.geoserver.wms.map.RenderedImageMapOutputFormat.produceMap(RenderedImageMapOutputFormat.java:569) [...] (I don't know if further trace is relevant, I think the problem is in ens ensureSortProperties) What's wrong with this style ? PS : how come I can't find any YSLD example anymore ? The YSLD reference does not mention the vendor options like x-sortBy anymore, is this format being abandonned ? Thanks a lot ! Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImagePyramid and TIFF overviews
Hi Daniele, thanks a lot for your help ! Le 25/10/2018 à 12:19, Daniele Romagnoli a écrit : Levels=0.2,0.2;0.4,0.4;0.8,0.8 1.6,1.6;3.2,3.2;6.4,6.4 12.8,12.8;25.6,25.6;51.2,51.2 102.4,102.4;204.8,204.8;409.6,409.6;819.2,819.2;1638.4,1638.4;3276.8,3276.8;6553.6,6553.6 LevelsNum=4 Envelope2D=-200,300 200,700 (I have copied your :-) ) LevelsDirs=0 1 2 3 Note the ";" char usage to specify available resolutions within the same pyramid level whilst the " " (whitespace) still separates different levels Excellent tip ! I am not sure I would have figured it out by myself ! :) Does it helps? A lot ! Thanks again ! Then I guess that you configure each level's "mosaic" in the same way, i.e. your "0" dir's property file looks like : Levels=0.2,0.2 0.4,0.4 0.8,0.8 (no ";" here I guess) Name=... [...] LevelsNum=3 Envelope2D=-200,300 200,700 Right ? Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] ImagePyramid and TIFF overviews
Hi all Is it possible to build an ImagePyramid with GeoTIFFs with overviews ? Actually, we have one raster folder per zoom level, and they are served as an ImagePyramid (index is in PostGIS). But since our dataset is updated very often, the process of recreating the rasters is long. Disk usage is also suboptimal. We have tried building GeoTIFF with overviews so that one TIFF could serve 2 to 3 levels (its a maximum because the displayed content changes every 2-3 zoom level) and are very satisfied with the result, so we would like to replace our "one folder per zoom level" structure to a "one folder per N zoom levels" structure with geotiff with 1 or 2 overviews inside. Is this possible ? If so, how should I configure the Imagepyramid properties file ? For the moment, we have something like : Name=georasters_pyramid Levels=0.15875,0.15875 0.3175,0.3175 0.635,0.635 1.27 1.27 LevelsDirs=georasters_17 georasters_16 georasters_15 georasters_14 LevelsNum=4 Envelope2D=-200,300 200,700 Thanks a lot ! Best regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] JAI merging granules performance
Hi Andrea, thanks for your answer ! Le 04/10/2018 à 13:05, Andrea Aime a écrit : do you have the "use JAI ImageRead" parameter turned on? (it is by default) In that case the read is deferred, happens when the JAI chain starts pulling tiles, so you see also the data reading as part of the image processing time. I have AllowMultiThread = true and USE_JAI_IMAGEREAD = false. I had seen this suggestion in many blogs / threads so it is the first one that I applied. Side note, LZW decompression is awfully slow, if you want to use lossless compression I'd suggest to switch to DEFLATE instead Thanks for the advice. Unfortunately it is an old dataset of nearly 1 million rasters, and the actual backend use to stich them together (closed source dll build on who knows what) does not support many options. I'm switching from that to GeoServer, but I'd like to keep the same performance. I already managed to get a small improvement by setting priority to 10 in JAI. Not the best, but OK anyway. Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] JAI merging granules performance
Hi all, I am running Geoserver 2.12.1 on Windows (Java8, 32bits, JAI & JAI ImageIO installed). I have set up an ImagePyramid wich works OK, but performance is not as good as what I expected. From the logs, the time taken between the parsing of the request and the end of the granules collection is OK (~ 200ms to query the index stored in a PostGIS DB, collect 6 granules, which are 1000x1000px TIFF files with LZW compression). But then, it takes almost 1 second for JAI to merge these granules into the final PNG. I have enabled multithreading on this layers which helped quite a lot, and disabled JAI_IMAGEREAD which helped a little. The log files shows nothing special happening, so I guess it's just the process that is not very fast : https://pastebin.com/EH8K6tUJ I have left the defaults in JAI and Coverage Access settings. Is there anything I could try to make things faster ? Thanks a lot ! -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] One of two overlapping polylines not rendered
For the sake of completeness : Le 20/09/2018 à 14:46, Arnaud L. a écrit : The layer was based on an SQL query with UNIONs. It "should" have been OK since these features where not real duplicates, but as I just found out the operator used for DISTINCT queries on geometries is the "=" operator, and this operator is not a real equality as is "st_equals". Replacing UNION with UNION ALL fixed the problem. As per the blog post by Paul Ramsey (http://blog.cleverelephant.ca/2017/09/postgis-operators.html), this was true only in PostGIS up to version 2.3. Starting with PostGIS 2.4, "=" is a real ST_Equals, and this should not occur anymore. Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] One of two overlapping polylines not rendered
Hi Andrea Le 20/09/2018 à 14:04, Andrea Aime a écrit : did you install the Marlin renderer? Or are you using the Windows installer? You can check in the status page if it's there (otherwise you'll get names like PiscesRenderer or DuctusRenderer): Selezione_387.png If you are using Marlin, one thing to try out is to disable it (how, depends on how you're running GeoServer). I just found the problem ! Contrary to what I thought, not all the features where in the result set. The layer was based on an SQL query with UNIONs. It "should" have been OK since these features where not real duplicates, but as I just found out the operator used for DISTINCT queries on geometries is the "=" operator, and this operator is not a real equality as is "st_equals". Replacing UNION with UNION ALL fixed the problem. This also explains why I had all the geometries in pgAdmin : I added the id field to identify the problematic geometries (stupid...). BTW, I am running in Windows but I already had the Marlin renderer installed. Thanks for your help ! Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] One of two overlapping polylines not rendered
Le 19/09/2018 à 12:04, Arnaud L. a écrit : I have a layer with closed polylines. Some of them almost fully overlap (but not quite). The problem I witness is that for some couple of polylines, only one of the two is rendered. Depending on the bbox, it can be one or the other, but never both. ... I manually ran both queries that I found in the log againt my PostGIS database, and both did return all the geometries (even the one not rendered). I checked in the KML layer preview and can confirm that everything is there. So the problem is probably somewhere in the rasterization process. Any idea ? Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] One of two overlapping polylines not rendered
Hi Olyster, and thanks for your answer. Le 19/09/2018 à 14:43, Olyster a écrit : I have encoutered this behavior and it was a problem with geometry type, SRID, bounding boxes. Verify eveything is fine in the layer config. I don't see what could be wrong in the layer config. All geometries are in the same PostGIS table and have the same SRID (EPSG:32631). The layer has the same SRID (EPSG:32631). One of the geometries is drawn and the other not, picking which one is drawn somehow randomely depending on BBOX. What could be wrong ? -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] One of two overlapping polylines not rendered
Hi all, I have a layer with closed polylines. Some of them almost fully overlap (but not quite). The problem I witness is that for some couple of polylines, only one of the two is rendered. Depending on the bbox, it can be one or the other, but never both. Here is an example of what I see : https://i.imgur.com/PZVNTOs.png https://i.imgur.com/d2VJ5IG.png As you can see, the second one is zoomed out and the polyline is shown. I over-simplified the style to rule this out, so it now reads : - name: 81 rules: - zoom: [13, 17] filter: ${layer IN ('81')} symbolizers: - line: uom: meter stroke-color: '#00' stroke-width: 0.1 I checked in the log (GEOTOOLS_DEVELOPPER_LOGGING level), but saw nothing wrong. I manually ran both queries that I found in the log againt my PostGIS database, and both did return all the geometries (even the one not rendered). The only thing that suprised me was : 2018-09-19 11:56:12,816 DEBUG [org.geotools.rendering] - Style cache hit ratio: 0.98 , hits 49, requests 50 And then : 2018-09-19 11:57:39,422 DEBUG [org.geotools.rendering] - Style cache hit ratio: 0.9655172413793104 , hits 28, requests 29 So in both cases, one hit less than requests. Does anyone have an idea of what could be happening here ? Thanks for helping ! Regards -- Arnaud ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] filter vs expression performance
Hi Jim Le 01-06-18 à 16:51, Jim Hughes a écrit : This is good question, and I'd be interested to know the answer. As a gut reaction, I don't think it'll matter too much. That said, there are several variables about this which will be specific to your use case. The best general advice is probably to set up a way to measure performance. GeoSolutions has free training materials which document using JMeter for testing just like this: https://geoserver.geo-solutions.it/edu/en/enterprise/jmeter.html. Thanks for the tips and the link. Actually, I did not have too much time to spend in this, so I've set up a very simple test case, running hundreds of curl commands against the same WMS url but with the two different styles. As you gyessed, there was evidence of a difference in rendering time. Since this will be a low-load service. I'm happy with this conclusion and I'll just use the syntax that offers the best readability depending on the case. Regards -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] filter vs expression performance
Hi list I'm wrinting a new style and I have to make choices between filters and expressions. I was wondering which one would be more efficient in terms of rendering performance : 2 rules with static style, or one rule with dynamic style ? One simple example would be (in ysld) : Filter version, two rules : rules: - filter: ${railway = 'rail' AND tunnel <> 'yes'} symbolizers: - line: stroke-color: '#99' rules: - filter: ${railway = 'rail' AND tunnel = 'yes'} symbolizers: - line: stroke-color: '#DD' Expression version, one rule, expression in style : rules: - filter: ${railway = 'rail'} symbolizers: - line: stroke-color: ${if_then_else(isNull(tunnel),'#99','#DD')} I have mixed feelings about both. I can write filter expressions more easily because the syntax is simple, and I am not familiar with CQL. But I like the compacity of the expression version better. Thanks for your thoughts ! -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] YSLD : expression in parameter expecting tupple
Le 23-04-18 à 15:52, Arnaud L. a écrit : I tried to recreate them from scratch, but I when trying to write an expression for a parameter expecting a tuple as input (namely, "anchor" and "displacement" parameters). How am I suppose to write this in ysld ? I found the answer in http://docs.geoserver.org/stable/en/user/styling/workshop/ysld/point.html The syntax is : ['${label}', ...] It is not very clear to me why this is the correct syntax, but it works. I guess I'll have to learn a bit more of yaml and ysld. -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] YSLD : expression in parameter expecting tupple
Hi all I am trying to convert some SLD styles to YSLD. When following the instructions for automatic conversion via the REST interface, I get errors (NullPointerException) for some styles. I tried to recreate them from scratch, but I when trying to write an expression for a parameter expecting a tuple as input (namely, "anchor" and "displacement" parameters). How am I suppose to write this in ysld ? In sld, it is as straightforward as with any other parameter, simply using propertyname where appropriate : anchx anchy But in YSLD, how am I supposed to write this ? I tried different bracket combinations, but none of them worked, so I thought maybe that was the reason for my conversion error : this could not be converted to YSLD ? Thanks for your help Regards -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImageMosaic stitching precision
Le 13/04/2017 à 11:49, Andrea Aime a écrit : > Ideally, try to setup a new mosaic with just 2 granules, with a > shapefile index, and see if you can reproduce the problem. > Increase the number of granules until you can, then pack everything into > a zip file and attach it to an issue in Jira. OK, I'll do that ASAP. Probably not this week though. > Just to be sure, are you reprojecting the data? If you are, it would be > sort of normal to have such jump every now and then. AFAIK, no. I'm trying very hard not to reproject, but indeed this does looks like a reprojection jump. The tif is in EPSG:32631 (but it has no .prf file, only a .tfw file), the imagemosaic is configured with native src EPSG:32631 and data src EPSG:32631. I am querying this layer via wms and specifying CRS=EPSG:32631. That's all I can think of. The only problem I see is that the tif have no prj files, but since the mosaic is defined with the correct data src I think it should not matter, right ? Regards -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImageMosaic stitching precision
Le 13/04/2017 à 10:09, Andrea Aime a écrit : > On Wed, Apr 12, 2017 at 2:12 PM, Arnaud L. <arnaud.lis...@codata.eu > <mailto:arnaud.lis...@codata.eu>> wrote: > > Trying to make a sample dataset I realized that this was not a > stitching problem, because the glitch is inside a tile, not on a > border between tiles. > So this is something totally different I believe... or is it not ? > > > It may be, hard to tell what is it without actually looking into it I'd be glad to help you with this. What should I provide to make it easy to replicate this ? I can provide the corresponding tiff with its world file, but will it be enough ? The index is in a postgis database, so I could dump the appropriate line ? Regards -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImageMosaic stitching precision
Le 12/04/2017 à 13:59, Andrea Aime a écrit : > That setting is used only while reprojecting, has no control over stitching. > For the residual issues, plan B, "open a ticket in the bug tracker with > sample data and > requests showing the issue, so that developers can debug it on their own > machine" Trying to make a sample dataset I realized that this was not a stitching problem, because the glitch is inside a tile, not on a border between tiles. So this is something totally different I believe... or is it not ? PS : Of course I checked the tile itself and it looks good, the glitch is only on the WMS representation. Regards -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImageMosaic stitching precision
Le 12/04/2017 à 10:35, Arnaud L. a écrit : > Le 12/04/2017 à 10:28, Andrea Aime a écrit : >> I've recently made a change that _might_ >> help: https://osgeo-org.atlassian.net/browse/GEOT-5615 >> The fix is present in GeoServer 2.10.2 and 2.11.0. If you are not using >> them, I'd suggest to upgrade and >> see if that helps, > > Indeed, I'm still on 2.8.4. I'll give 2.11.0 a try ASAP, hoping that the > upgrade will go smoothly. 2.11.0 does help, thanks ! I don't have black lines anymore. I still do have some stitching problems though, as illustrated here : http://i.imgur.com/0VUlYWU.png I tries with both "accurate resolution computation" on and off, but this setting has no effect on the problem. BTW, 2.8.4 was running without JAI (it was a mistake, because of an update of the JRE), and 2.11.0 is running with JAI. I don't know if this alone could explain the difference. Also I am running with multithreading on and therefore JAI ImageRead off, so I'm not sure that JAI is used anyway. Regards -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
Re: [Geoserver-users] ImageMosaic stitching precision
Hi Aime, thanks a lot for the quick answer ! Le 12/04/2017 à 10:28, Andrea Aime a écrit : > I've recently made a change that _might_ > help: https://osgeo-org.atlassian.net/browse/GEOT-5615 > The fix is present in GeoServer 2.10.2 and 2.11.0. If you are not using > them, I'd suggest to upgrade and > see if that helps, Indeed, I'm still on 2.8.4. I'll give 2.11.0 a try ASAP, hoping that the upgrade will go smoothly. I'll keep you informed, and I'll have a look at the tracker anyway. Thanks again ! -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] ImageMosaic stitching precision
Hi all, I have set up an ImageMosaic store with existing TIFF tiles. They are not GeoTIFFs, but regular TIFF files with worldfiles. The index is stored in PostGIS (the information was already present there, I only had to create a view to match the index schema). In some cases, the stitching of rasters is not perfect and a black line appears between tiles. The rasters are in EPSG:32631, and I am also querying GeoServer's WMS in EPSG:32631, so no reprojection is done. The rasters are all 1000 pixels in widht and height, the resolution is 0.15875 (in EPSG 32631, so 0.15875 m/pix), so each raster is 158.75m in width and height. Here is an example request, and the resulting raster : WMS url : http://localhost/geoserver/dev/wms?SERVICE=WMS=1.3.0=GetMap=image%2Fpng=true=dev%3Ageorasters_test=EPSG%3A32631=1920=1349=742405.64625%2C5087266.7415625%2C742710.44625%2C5087480.895312501 Raster : http://i.imgur.com/pE2H4QJ.png Slightly changing the bbox fixes things up (1 meter to the east) : WMS url : http://localhost/geoserver/dev/wms?SERVICE=WMS=1.3.0=GetMap=image%2Fpng=true=dev%3Ageorasters_test=EPSG%3A32631=1920=1349=742406.64625%2C5087266.741562501%2C742711.44625%2C5087480.895312501 Raster : http://i.imgur.com/PjflMLd.png In my GeoServer layer's configuration, I have tried with "Accurate resolution computation" to true and false, to no avail (both examples above are with this setting set to true). Is there any other setting controlling the stitching mechanism ? Where could be the problem here ? In the index, or in the data ? Both are supposed to be 100% accurate, and indeed the both work without problem with another stitching library (no idea what it is exactly, it's a custom dll developed by a third party in .NET). Thanks a lot for your help ! -- Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users
[Geoserver-users] ImageMosaic with predefined projection
Hi all, I'm migrating a bunch of our services and data to GeoServer (2.8.4, Windows). We have a quite large raster store here. It consists of some hundreds of thousands of TIFF files with their accompanying TFW worldfile. They are all in the same projection. I thought it would be easy to configure an ImageMosaic with this data since all the needed information is already stored in a PostGIS database, but I have an issue because our TIFFs are not GeoTIFFs, and they also don't have a PRJ file. I could create a PRJ file for every TIFF, but that would mean a lot of wasted space (and time). I could also convert all of them to GeoTIFF, but that's a hard work. I have yet to find a tool that *modifies* a TIFF to make it a GeoTIFF. All the tools I've found make copy of images (i.e. create a new GeoTIFF file instead of converting the TIFF to GeoTIFF). This does not integrate well in our processes at all. Is there a way to setup my datastore so that GeoServer doesn't look for a PRJ file and just uses a projection that I would hardcode somewhere ? Thanks a lot for your help ! Arnaud -- Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot ___ Geoserver-users mailing list Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users