Upgrade solr to 9.10.1 - #503
Conversation
VictorHarbo
added
dependencies
There was a problem hiding this comment.
1 bug that must be fixed, and try research if you do avoid hardcoded magic strings.
Press F5 after a search to try again and the get the error.
2026-06-24 07:23:23 [http-nio-8080-exec-13] ERROR dk.kb.netarchivesuite.solrwayback.service.SolrWaybackResourceWeb(SolrWaybackResourceWeb.java:676) - ServiceException(HTTP 500) in Service:
java.io.IOException: closed
at java.net.http/jdk.internal.net.http.ResponseSubscribers$HttpResponseInputStream.current(ResponseSubscribers.java:352)
at java.net.http/jdk.internal.net.http.ResponseSubscribers$HttpResponseInputStream.read(ResponseSubscribers.java:403)
at java.base/java.io.InputStream.readNBytes(InputStream.java:396)
at java.base/java.io.InputStream.readAllBytes(InputStream.java:333)
at dk.kb.netarchivesuite.solrwayback.solr.NetarchiveSolrClient.requestRawJson(NetarchiveSolrClient.java:1542)
at dk.kb.netarchivesuite.solrwayback.solr.NetarchiveSolrClient.searchJsonResponseOnlyFacets(NetarchiveSolrClient.java:1419)
at dk.kb.netarchivesuite.solrwayback.facade.Facade.proxySolrOnlyFacets(Facade.java:1132)
| NamedList<Object> resp = solrServer.request(req); | ||
| String jsonResponse = resp.get("response").toString(); | ||
| return jsonResponse; | ||
| int status = (int) resp.get("responseStatus"); |
There was a problem hiding this comment.
It is this not possible without hardcoded strings? Is there Constants defined somewhere?
There was a problem hiding this comment.
The cause of the bug is somehow related to these new lines
There was a problem hiding this comment.
Found the constants and have resolved the bug.
|
Good catch @thomasegense! I will look into it! |
thomasegense
left a comment
thomasegense
left a comment
There was a problem hiding this comment.
LGTM.
The custom made caching framework was quite a trap and had nothing to do with the suprisingly easy solr upgrade.
I am not that pleased with the caching/superclasses to solrclient, too complicated.
2 participants
This pull request updates Solr and Lucene dependencies and refactors how raw JSON responses are retrieved from Solr in the
NetarchiveSolrClientclass. The refactor replaces the previous use ofNoOpResponseParserwith a new method leveragingInputStreamResponseParser, as NoOpResponseParser is deprecated.Dependency updates:
pom.xmlto newer versionsRefactoring JSON response handling:
NoOpResponseParserwithInputStreamResponseParser, as the first one is deprecatedPLEASE HAVE A LOOK at the use of InputStreamResponseParser, as I am using the stream.readAllBytes() which is not good for large amounts of data, but I think these queries are only used for returning search results right?
I've tested in the following way:
Ps. This has not been reviewed internally in the WEBCHILD team.