Incident #9908
closed
geoanalytics-administration portlet not working
100%
Description
The portlet reports "Failed to contact server. Maybe server is offline"
https://bluebridge.d4science.org/group/siaspa/geoanalytics-administration
and the only thing it logs is this:
[PORTAL] 1101829 [http-bio-9090-exec-64] INFO gr.cite.bluebridge.endpoint.EndpointManager - Returned 1 service endpoint(s)
I'll be travelling on Monday, @costantino.perciante@isti.cnr.it can give you support.
Updated by Massimiliano Assante over 7 years ago
- Tracker changed from Support to Incident
Updated by Dimitris Katris over 7 years ago
- Status changed from New to In Progress
In order to solve this issue we had to revert a change that was performed in the portlet recently. This change was creating a problem that was only visible in production environment. When the portlet was deployed in the development or in the pre-production it was working without any problem.
I will try to give a brief explanation regarding this issue. In general, the geoanalytics administration portlet performs only some simple actions i.e. discovering the active scope, user, active service instance, and afterwards it dispatches the request to the geoanalytics service that does the actual work. Furthermore, the response that the client (user's browser) receives is actually the response that the service generates. In the latest portlet update apart, from the response body we also changed the code in order to return also all the http headers that the service creates. By doing this, the portlet probably overrides some http headers that are required from the production portal to work properly. After we reverted this change, the portlet seems to work without any issues.
A new portlet configuration has been created (#9914) in order to fix this issue.
Dimitris
Updated by Massimiliano Assante over 7 years ago
Thank you for the explanation Dimitris, however I would like to understand better what is this difference between the production and pre production portal you are talking about. The aim is to have the exact environment in both portals so that this kind of issue would emerge before and not after putting the portlet in production. You mentioned the portlet probably overrides some http headers that are required from the production portal to work properly. I would like to know more about this, perhaps we can involve our sysadmins in this.
Updated by Dimitris Katris over 7 years ago
@massimiliano.assante@isti.cnr.it Can we close this incident and mark it as resolved? The process of investigating the difference between the production and pre-production could be continued with the help of @vfloros@cite.gr in the background or by creating a new ticket. But I think that this could be closed now since the issue is now fixed.
Updated by Massimiliano Assante over 7 years ago
- Status changed from In Progress to Closed
- % Done changed from 0 to 100
Yes, absolutely. I am closing the incident, please create a new ticket with the investigation (if not already open).
