Bug #18690
closedLinks still pointing to BlueBRIDGE
100%
Description
As @anton.ellenbroek@fao.org noticed, links in exploiting fishery and similar records are still pointing to bluebridge, please see attached screeshot (from this record http://data.d4science.org/ctlg/GRSF/f3849a07-7c11-3206-bf5c-06199b703675)
Files
Updated by Francesco Mangiacrapa about 5 years ago
- Status changed from New to In Progress
Updated by Francesco Mangiacrapa about 5 years ago
- Status changed from In Progress to Feedback
- % Done changed from 0 to 100
Hi @aureliano.gentile@fao.org,
I just updated the configurations to resolve also the links of GRSF_ADMIN (e.g. this one http://data.d4science.org/ctlg/GRSF_Admin/5d609960-47b1-3d81-89c0-1373bc0fd4dc) towards the iMarine-Gateway. Let me know...
Updated by Francesco Mangiacrapa about 5 years ago
- Status changed from Feedback to Closed
Updated by Aureliano Gentile about 5 years ago
Dear Francesco, right this morning I was checking this ticket (my apologies for the delay).
I confirm it is now fine in the GRSF Admin.
Regarding the public release https://i-marine.d4science.org/web/grsf/data-catalogue I made a search by "bluebridge" and I got this record (Makaira nigricans Atlantic)
which indeed contains this url
not sure why...
Updated by Francesco Mangiacrapa about 5 years ago
- Status changed from Closed to In Progress
- % Done changed from 100 to 80
Hi Aureliano,
I'm adding @marketak@ics.forth.gr to this ticket.
Regarding the issue reported in your previous comment (at 18690#note-4), in my opinion during the publishing phase these links have been resolved on KB side from RECORD URL of kind: http://data.d4science.org/ctlg/GRSF/[UUID] to resolved link like: https://bluebridge.d4science.org/group/grsf_admin/data-catalogue?path%3D%2Fdataset%2F[UUID] and published into GRSF/GRSF_ADMIN Catalogue as link already resolved instead of RECORD URL.
Only the first type of link (the RECORD URL) must be stored into Catalogue not the second one (the link resolved).
The RECORD URL (e.g. http://data.d4science.org/ctlg/GRSF/e9d96077-6cb7-34c2-bf06-dc04e25dad89), is resolved dynamically by URI-Resolver service at consuming-time (e.g. by pasting the link into browser). With URI-Resolver service we are able to resolve a RECORD URL (in general an ITEM URL published into a D4Science Catalogue) always pointing to the current gateway/catalogue that contains and provides the RECORD in the D4Science Infrastructure. In fact, just changing one configuration (on Information System-side) we have been able to switch from bluebridge.d4science.org to i-marine.d4science.org by using the (same) RECORD URLs already published in the GRSF/GRSF_ADMIN Catalogues.
Instead, the link resolved (e.g. https://bluebridge.d4science.org/group/grsf_admin/data-catalogue?path%3D%2Fdataset%2F7febfe97-2952-3392-bde2-dfc48d754ff5) already points to destination and it is not dynamic.
Regarding the GRSF public release at https://i-marine.d4science.org/web/grsf/data-catalogue
Searching for "bluebridge.d4science" only 1 record is found http://data.d4science.org/ctlg/GRSF/e9d96077-6cb7-34c2-bf06-dc04e25dad89
and it contains a 'Similar GRSF Record' with url https://bluebridge.d4science.org/group/grsf_admin/data-catalogue?path%3D%2Fdataset%2F7febfe97-2952-3392-bde2-dfc48d754ff5
Regarding the GRSF_ADMIN release at https://i-marine.d4science.org/group/grsf_admin/data-catalogue
Searching for "bluebridge.d4science" I got 27 records (you can see them at https://i-marine.d4science.org/group/grsf_admin/data-catalogue?path=dataset&query=cT0iYmx1ZWJyaWRnZS5kNHNjaWVuY2Ui)
Yannis, please, could you republish them by fixing the RECORD URL?
Updated by Yannis Marketakis about 5 years ago
The URLs are not generated in the GRSF KB during the publishing. The actual workflow is:
- We publish a record X (which does not have a URL yet)
- As soon as it is published in the catalogue, the publishing client receives the catalog ID and the catalogue URL
- The catalogue URL is added in GRSF KB
- If necessary (e.g. if there are similar records) we update records so that their URLs are visible in the catalogue as well.
Now the reason for having URLs with two different prefixes is because some of them have been published a long time ago (when a particular prefix was used), while others are more recent (with the new prefix).
I guess I could update the prefixes of the URLs in the GRSF KB (manually) and update the affected records. Is that fine?
Updated by Francesco Mangiacrapa about 5 years ago
Yannis Marketakis wrote:
The URLs are not generated in the GRSF KB during the publishing. The actual workflow is:
- We publish a record X (which does not have a URL yet)
- As soon as it is published in the catalogue, the publishing client receives the catalog ID and the catalogue URL
- The catalogue URL is added in GRSF KB
- If necessary (e.g. if there are similar records) we update records so that their URLs are visible in the catalogue as well.
Many thanks @marketak@ics.forth.gr for clarifying the workflow.
Now the reason for having URLs with two different prefixes is because some of them have been published a long time ago (when a particular prefix was used), while others are more recent (with the new prefix).
I guess I could update the prefixes of the URLs in the GRSF KB (manually) and update the affected records. Is that fine?
If for updating the prefixes of the URLs in the GRSF KB (manually) ... you mean to use the RECORD URL of kind:
- for GRSF http://data.d4science.org/ctlg/GRSF/[THE_RECORD_UUID] (e.g. http://data.d4science.org/ctlg/GRSF/e9d96077-6cb7-34c2-bf06-dc04e25dad89)
- for GRSF_ADMIN http://data.d4science.org/ctlg/GRSF_ADMIN/[THE_RECORD_UUID] (e.g. http://data.d4science.org/ctlg/GRSF_Admin/255b8039-dfc8-3598-954e-2ad5fbf95f7e)
that's OK.
Otherwise, you can obtain programmatically the Record URL by performing a POST operation to http://data.d4science.org/ctlg
(you can see how to at https://wiki.gcube-system.org/gcube/URI_Resolver#CATALOGUE_Resolver)
Updated by Yannis Marketakis about 5 years ago
Francesco Mangiacrapa wrote:
If for updating the prefixes of the URLs in the GRSF KB (manually) ... you mean to use the RECORD URL of kind:
- for GRSF http://data.d4science.org/ctlg/GRSF/[THE_RECORD_UUID] (e.g. http://data.d4science.org/ctlg/GRSF/e9d96077-6cb7-34c2-bf06-dc04e25dad89)
- for GRSF_ADMIN http://data.d4science.org/ctlg/GRSF_ADMIN/[THE_RECORD_UUID] (e.g. http://data.d4science.org/ctlg/GRSF_Admin/255b8039-dfc8-3598-954e-2ad5fbf95f7e)
that's OK.
That's exactly what I mean. So, I'll proceed with this.
Updated by Francesco Mangiacrapa about 5 years ago
- Assignee changed from Francesco Mangiacrapa to Yannis Marketakis
Yannis Marketakis wrote:
Francesco Mangiacrapa wrote:
If for updating the prefixes of the URLs in the GRSF KB (manually) ... you mean to use the RECORD URL of kind:
- for GRSF http://data.d4science.org/ctlg/GRSF/[THE_RECORD_UUID] (e.g. http://data.d4science.org/ctlg/GRSF/e9d96077-6cb7-34c2-bf06-dc04e25dad89)
- for GRSF_ADMIN http://data.d4science.org/ctlg/GRSF_ADMIN/[THE_RECORD_UUID] (e.g. http://data.d4science.org/ctlg/GRSF_Admin/255b8039-dfc8-3598-954e-2ad5fbf95f7e)
that's OK.
That's exactly what I mean. So, I'll proceed with this.
Many Thanks
Updated by Yannis Marketakis about 5 years ago
@francesco.mangiacrapa@isti.cnr.it I found 4 defected records using the old (bluebridge-like URI).
I've updated them manually, however I noticed that after updating them in the catalogue, GRSF Publisher service is still returning me the old URL (see log attached).
Is there any issue on the publisher service side? or should I just try to delete and re-publish the record (instead of updating it)?
Updated by Yannis Marketakis about 5 years ago
- File updateLog.log updateLog.log added
Updated by Luca Frosini about 5 years ago
- Assignee changed from Yannis Marketakis to Luca Frosini
Updated by Luca Frosini about 5 years ago
- Assignee changed from Luca Frosini to Francesco Mangiacrapa
I have found this exception:
2020-03-26 07:57:31,197 [Thread-40] ERROR JCRWorkspaceItem: Impossible to get short url. Long url will be returned. 2020-03-26 07:57:33,867 [Thread-52] WARN UrlShortener: Runtime Resource HTTP-URL-Shortener-DL not found 2020-03-26 07:57:33,867 [Thread-52] ERROR UrlShortener: An error occurred reading Runtime Resource for name: HTTP-URL-Shortener-DL, the scope is: /d4science.research-infrastructures.eu/FARM/GRSF_Admin java.lang.Exception: No Runtime Resource with name: HTTP-URL-Shortener-DL is available in the scope: /d4science.research-infrastructures.eu/FARM/GRSF_Admin at org.gcube.portlets.user.urlshortener.UrlShortener.<init>(UrlShortener.java:97) at org.gcube.portlets.user.urlshortener.UrlShortener.<init>(UrlShortener.java:69) at org.gcube.common.homelibrary.jcr.workspace.JCRWorkspaceItem.getShortUrl(JCRWorkspaceItem.java:1189) at org.gcube.common.homelibrary.jcr.workspace.JCRWorkspaceItem.getPublicLink(JCRWorkspaceItem.java:1146) at org.gcube.data_catalogue.grsf_publish_ws.utils.csv.ManageTimeSeriesThread.manageTimeSeries(ManageTimeSeriesThread.java:215) at org.gcube.data_catalogue.grsf_publish_ws.utils.csv.ManageTimeSeriesThread.run(ManageTimeSeriesThread.java:90)
@francesco.mangiacrapa@isti.cnr.it is not working today. We have to wait for him.
Updated by Luca Frosini about 5 years ago
@marketak@ics.forth.gr can you please provide me with the JSON you are using to update the record so that I can made some test?
Updated by Luca Frosini about 5 years ago
I @marketak@ics.forth.gr
thanks to your and @francesco.mangiacrapa@isti.cnr.it suggestion I found that the Record URL is explicitly kept as is.
Is it a problem for you to delete and recreate the records?
Updated by Luca Frosini about 5 years ago
- Assignee changed from Francesco Mangiacrapa to Luca Frosini
Updated by Yannis Marketakis about 5 years ago
I guess I could try deleting and republishing.
I'll let you know.
Updated by Yannis Marketakis about 5 years ago
I've deleted and re-published 4 stock records. Now their URLs are OK.
In addition, I've updated several stock and fishery records that were linked to those 4 (either as similar records or as exploited resources). Now they seem OK.
Updated by Aureliano Gentile about 5 years ago
with many thanks, in the mean time I am checking the new data harvest, I assume all of these fixes will be inherited.
Updated by Luca Frosini about 5 years ago
- Status changed from In Progress to Closed
- % Done changed from 80 to 100