Re: Solr alternates returning different versions of the same document
Hi Erik, thanks - if it helps, I eventually fixed the problem by deleting the documents by id (via an http request), which apparently deleted all the versions everywhere, then re-creating the documents via the admin interface (update, csv). This seems to have left only one version of each document. Yann -- View this message in context: http://lucene.472066.n3.nabble.com/Solr-alternates-returning-different-versions-of-the-same-document-tp4143006p4143680.html Sent from the Solr - User mailing list archive at Nabble.com.
Re: Solr alternates returning different versions of the same document
Thanks for letting us know. Erick On Tue, Jun 24, 2014 at 5:25 AM, yann yannick.lallem...@gmail.com wrote: Hi Erik, thanks - if it helps, I eventually fixed the problem by deleting the documents by id (via an http request), which apparently deleted all the versions everywhere, then re-creating the documents via the admin interface (update, csv). This seems to have left only one version of each document. Yann -- View this message in context: http://lucene.472066.n3.nabble.com/Solr-alternates-returning-different-versions-of-the-same-document-tp4143006p4143680.html Sent from the Solr - User mailing list archive at Nabble.com.
Re: Solr alternates returning different versions of the same document
Hi Erik, thanks for your answer. I didn't manually assign docs to shards, I indexed all docs on one server, which then assigned it to shards (based on the default Solr behaviour, based on the document ID I believe). If I understood you correctly - this means the update section of the admin should be avoided when using a sharded install, because it doesn't guarantee a given document ID will be sent to the same shard as the previous version of the same document? Thanks Yann -- View this message in context: http://lucene.472066.n3.nabble.com/Solr-alternates-returning-different-versions-of-the-same-document-tp4143006p4143375.html Sent from the Solr - User mailing list archive at Nabble.com.
Re: Solr alternates returning different versions of the same document
bq: If I understood you correctly - this means the update section of the admin should be avoided when using a sharded install, because it doesn't guarantee a given document ID will be sent to the same shard as the previous version of the same document? You've got it, but I want to emphasize that I'm _speculating_ here based on the fact that the admin pages originated in non-solr-cloud days and are fairly cloud-ignorant. Someone would have to, I think, do something special to make indexing via the admin UI cloud-aware. So my scenario seems at least plausible... Hmmm, this is worth a JIRA though, I've raised one. P.S. There's the bare start of an umbrella project of things to do for a SolrCloud-aware admin UI here: https://issues.apache.org/jira/browse/SOLR-6082 Best, Erick On Mon, Jun 23, 2014 at 12:42 AM, yann yannick.lallem...@gmail.com wrote: Hi Erik, thanks for your answer. I didn't manually assign docs to shards, I indexed all docs on one server, which then assigned it to shards (based on the default Solr behaviour, based on the document ID I believe). If I understood you correctly - this means the update section of the admin should be avoided when using a sharded install, because it doesn't guarantee a given document ID will be sent to the same shard as the previous version of the same document? Thanks Yann -- View this message in context: http://lucene.472066.n3.nabble.com/Solr-alternates-returning-different-versions-of-the-same-document-tp4143006p4143375.html Sent from the Solr - User mailing list archive at Nabble.com.
Solr alternates returning different versions of the same document
I have the following problem with Solr 4.5.1, with a cloud install with 4 shards, no replication, using the built-in zookeeper on one Solr: I have updated a document via the Solr console (select a core, then select Documents). I used the CSV format to upload the document, including the document ID. When I query the document id from the Solr console (simple query: id:the-id-of-the-doc-I-updated), I alternatively obtain the old document (with the values before update, and a given _version_ number), or the new document (with the values after update, and a different _version_). No log messages in the Solr console about updating the document or anything. Any idea what might be going on, and how to fix that problem? Thanks in advance, Yann -- View this message in context: http://lucene.472066.n3.nabble.com/Solr-alternates-returning-different-versions-of-the-same-document-tp4143006.html Sent from the Solr - User mailing list archive at Nabble.com.
Re: Solr alternates returning different versions of the same document
If you update to a specific core, I suspect you're getting the doc indexed on two shards which leads to duplicate documents being returned. So it depends on which core happens to answer the request... Fundamentally, all versions of a document must go to the same shard in order for the new version to replace the old version. If you've put the document specifically on a single node, you've bypassed the automatic routing that would insure this... I think the Admin UI kind of side-steps the usual routing process, but I'm not entirely sure. Best, Erick On Fri, Jun 20, 2014 at 12:47 AM, yann yannick.lallem...@gmail.com wrote: I have the following problem with Solr 4.5.1, with a cloud install with 4 shards, no replication, using the built-in zookeeper on one Solr: I have updated a document via the Solr console (select a core, then select Documents). I used the CSV format to upload the document, including the document ID. When I query the document id from the Solr console (simple query: id:the-id-of-the-doc-I-updated), I alternatively obtain the old document (with the values before update, and a given _version_ number), or the new document (with the values after update, and a different _version_). No log messages in the Solr console about updating the document or anything. Any idea what might be going on, and how to fix that problem? Thanks in advance, Yann -- View this message in context: http://lucene.472066.n3.nabble.com/Solr-alternates-returning-different-versions-of-the-same-document-tp4143006.html Sent from the Solr - User mailing list archive at Nabble.com.
