Re: newbie question on how to batch commit documents
: CommonsHttpSolrServer.request() resulting in multiple searchers. My first : thought was to change the configs for autowarming. But after looking at the : autowarm params, I am not sure what can be changed or perhaps a different : approach is recommened. even with 0 autowarming (which is what you have) it can still take time to close/open a searcher on every commit -- which is why a commit per doc is not usually a good idea (and is *definitely* not a good idea when doing batch indexing. most people can get away with just doing one commit after all their docs have been added (ie: at the end of the batch) but if you've got a ot of distinct clients, doing a lot of parllel indexing and you don't want to coordinate who is responsible for sending the commit, you can configure "autocommit" to happen on the solr server... http://wiki.apache.org/solr/SolrConfigXml#Update_Handler_Section ...but in general you should make sure that your clients sending docs can deal with the occasional long delays (or possibly even needing to retry) when an occasional commit might block add/delete operations because of an expensive segment merge. -Hoss
Re: newbie question on how to batch commit documents
I would additionally suggest to use embeddedSolrServer for large uploads if
possible, performance are better.
2010/5/31 Steve Kuo
> I have a newbie question on what is the best way to batch add/commit a
> large
> collection of document data via solrj. My first attempt was to write a
> multi-threaded application that did following.
>
> Collection docs = new ArrayList();
> for (Widget w : widges) {
>doc.addField("id", w.getId());
>doc.addField("name", w.getName());
> doc.addField("price", w.getPrice());
>doc.addField("category", w.getCat());
>doc.addField("srcType", w.getSrcType());
>docs.add(doc);
>
>// commit docs to solr server
>server.add(docs);
>server.commit();
> }
>
> And I got this exception.
>
> rg.apache.solr.common.SolrException:
>
> Error_opening_new_searcher_exceeded_limit_of_maxWarmingSearchers2_try_again_later
>
>
> Error_opening_new_searcher_exceeded_limit_of_maxWarmingSearchers2_try_again_later
>
>at
> org.apache.solr.client.solrj.impl.CommonsHttpSolrServer.request(CommonsHttpSolrServer.java:424)
>at
> org.apache.solr.client.solrj.impl.CommonsHttpSolrServer.request(CommonsHttpSolrServer.java:243)
>at
> org.apache.solr.client.solrj.request.AbstractUpdateRequest.process(AbstractUpdateRequest.java:105)
>at
> org.apache.solr.client.solrj.SolrServer.commit(SolrServer.java:86)
>
> The solrj wiki/documents seemed to indicate that because multiple threads
> were calling SolrServer.commit() which in term called
> CommonsHttpSolrServer.request() resulting in multiple searchers. My first
> thought was to change the configs for autowarming. But after looking at
> the
> autowarm params, I am not sure what can be changed or perhaps a different
> approach is recommened.
>
> class="solr.FastLRUCache"
> size="512"
> initialSize="512"
> autowarmCount="0"/>
>
> class="solr.LRUCache"
> size="512"
> initialSize="512"
> autowarmCount="0"/>
>
> class="solr.LRUCache"
> size="512"
> initialSize="512"
> autowarmCount="0"/>
>
> Your help is much appreciated.
>
Re: newbie question on how to batch commit documents
Add commit after the loop. I would advise to use commit in a separate
thread. I do keep separate timer thread, where every minute I will do
commit and at the end of every day I will optimize the index.
Regards
Aditya
www.findbestopensource.com
On Tue, Jun 1, 2010 at 2:57 AM, Steve Kuo wrote:
> I have a newbie question on what is the best way to batch add/commit a
> large
> collection of document data via solrj. My first attempt was to write a
> multi-threaded application that did following.
>
> Collection docs = new ArrayList();
> for (Widget w : widges) {
>doc.addField("id", w.getId());
>doc.addField("name", w.getName());
> doc.addField("price", w.getPrice());
>doc.addField("category", w.getCat());
>doc.addField("srcType", w.getSrcType());
>docs.add(doc);
>
>// commit docs to solr server
>server.add(docs);
>server.commit();
> }
>
> And I got this exception.
>
> rg.apache.solr.common.SolrException:
>
> Error_opening_new_searcher_exceeded_limit_of_maxWarmingSearchers2_try_again_later
>
>
> Error_opening_new_searcher_exceeded_limit_of_maxWarmingSearchers2_try_again_later
>
>at
> org.apache.solr.client.solrj.impl.CommonsHttpSolrServer.request(CommonsHttpSolrServer.java:424)
>at
> org.apache.solr.client.solrj.impl.CommonsHttpSolrServer.request(CommonsHttpSolrServer.java:243)
>at
> org.apache.solr.client.solrj.request.AbstractUpdateRequest.process(AbstractUpdateRequest.java:105)
>at
> org.apache.solr.client.solrj.SolrServer.commit(SolrServer.java:86)
>
> The solrj wiki/documents seemed to indicate that because multiple threads
> were calling SolrServer.commit() which in term called
> CommonsHttpSolrServer.request() resulting in multiple searchers. My first
> thought was to change the configs for autowarming. But after looking at
> the
> autowarm params, I am not sure what can be changed or perhaps a different
> approach is recommened.
>
> class="solr.FastLRUCache"
> size="512"
> initialSize="512"
> autowarmCount="0"/>
>
> class="solr.LRUCache"
> size="512"
> initialSize="512"
> autowarmCount="0"/>
>
> class="solr.LRUCache"
> size="512"
> initialSize="512"
> autowarmCount="0"/>
>
> Your help is much appreciated.
>
Re: newbie question on how to batch commit documents
Move the commit outside your loop and you'll be in better shape.
Better yet, enable autocommit in solrconfig.xml and don't commit from
your multithreaded client, otherwise you still run the risk of too
many commits happening concurrently.
Erik
On May 31, 2010, at 5:27 PM, Steve Kuo wrote:
I have a newbie question on what is the best way to batch add/commit
a large
collection of document data via solrj. My first attempt was to
write a
multi-threaded application that did following.
Collection docs = new
ArrayList();
for (Widget w : widges) {
doc.addField("id", w.getId());
doc.addField("name", w.getName());
doc.addField("price", w.getPrice());
doc.addField("category", w.getCat());
doc.addField("srcType", w.getSrcType());
docs.add(doc);
// commit docs to solr server
server.add(docs);
server.commit();
}
And I got this exception.
rg.apache.solr.common.SolrException:
Error_opening_new_searcher_exceeded_limit_of_maxWarmingSearchers2_try_again_later
Error_opening_new_searcher_exceeded_limit_of_maxWarmingSearchers2_try_again_later
at
org
.apache
.solr
.client
.solrj.impl.CommonsHttpSolrServer.request(CommonsHttpSolrServer.java:
424)
at
org
.apache
.solr
.client
.solrj.impl.CommonsHttpSolrServer.request(CommonsHttpSolrServer.java:
243)
at
org
.apache
.solr
.client
.solrj
.request.AbstractUpdateRequest.process(AbstractUpdateRequest.java:105)
at org.apache.solr.client.solrj.SolrServer.commit(SolrServer.java:86)
The solrj wiki/documents seemed to indicate that because multiple
threads
were calling SolrServer.commit() which in term called
CommonsHttpSolrServer.request() resulting in multiple searchers. My
first
thought was to change the configs for autowarming. But after
looking at the
autowarm params, I am not sure what can be changed or perhaps a
different
approach is recommened.
Your help is much appreciated.
newbie question on how to batch commit documents
I have a newbie question on what is the best way to batch add/commit a large
collection of document data via solrj. My first attempt was to write a
multi-threaded application that did following.
Collection docs = new ArrayList();
for (Widget w : widges) {
doc.addField("id", w.getId());
doc.addField("name", w.getName());
doc.addField("price", w.getPrice());
doc.addField("category", w.getCat());
doc.addField("srcType", w.getSrcType());
docs.add(doc);
// commit docs to solr server
server.add(docs);
server.commit();
}
And I got this exception.
rg.apache.solr.common.SolrException:
Error_opening_new_searcher_exceeded_limit_of_maxWarmingSearchers2_try_again_later
Error_opening_new_searcher_exceeded_limit_of_maxWarmingSearchers2_try_again_later
at
org.apache.solr.client.solrj.impl.CommonsHttpSolrServer.request(CommonsHttpSolrServer.java:424)
at
org.apache.solr.client.solrj.impl.CommonsHttpSolrServer.request(CommonsHttpSolrServer.java:243)
at
org.apache.solr.client.solrj.request.AbstractUpdateRequest.process(AbstractUpdateRequest.java:105)
at org.apache.solr.client.solrj.SolrServer.commit(SolrServer.java:86)
The solrj wiki/documents seemed to indicate that because multiple threads
were calling SolrServer.commit() which in term called
CommonsHttpSolrServer.request() resulting in multiple searchers. My first
thought was to change the configs for autowarming. But after looking at the
autowarm params, I am not sure what can be changed or perhaps a different
approach is recommened.
Your help is much appreciated.
