How to disable cache for facet.query?

[Wei]

Hi,

I am trying to disable filter cache for some filter queries as they contain
unique ids and cause cache evictions. By adding {!cache=false} the fq is no
longer stored in filter cache, however I have similar conditions in
facet.query and using facet.query={!cache=false}(color:red AND id:XXX) does
not work.  Is it possible to stop solr from putting facet.query into filter
cache?

Thanks,
Wei

Can create collections with Drupal 8 configset

[Shane Brooks]

Hello,

Main info: Drupal 8, SOLRCloud 7.7.3, Zookeeper 3.4.14, search_api 8.x-1.17, 
search_api_solr 4.14

Migrating from a standalone SOLR 7.7.3 instance where everything works as 
expected.

Drupal 7 configset gives us no issue on SOLRCloud.

Issue:
When I try to create a new collection using the Drupal 8 configset via POST:

http://solr.ourcompany.com:8983/solr/admin/collections?action=CREATE=ourcollection=1=2=true=d8configset=replica:<2,node:*

I get the following error:

org.apache.solr.client.solrj.impl.HttpSolrClient$RemoteSolrException:Error from 
server at http://192.168.xx.xx:8983/solr: Expected mime type 
application/octet-stream but got text/html. \n\n\nError 
500 Server Error\n\nHTTP ERROR 500\nProblem 
accessing /solr/admin/cores. Reason:\n Server ErrorCaused 
by:java.lang.NoClassDefFoundError: 
org/apache/lucene/collation/ICUCollationKeyAnalyzer\n\tat

I haven't found anything on Google except that ICUCollationKeyAnalyzer depends 
on the icu4j library, which I verified is part of the SOLR package.

Can anyone help me with this issue? Is there anything else I can provide?

Thanks,
Shane






Shane Brooks
Systems Administrator
Phone 800.262.2633 x 7465
Email sh...@municode.com
[A close up of a sign  Description automatically generated]
Connecting you and your community
through the circle of governance:

< Meeting Management

< Website Design

< Policy & Procedure Software

< Legislative Publishing Software

< Document Archival Tool

< Codification 
Services

Re: solr startup

[Erick Erickson]

It totally depends on how often you open a new searcher. If you have an index 
that updates once a day, running extensive autowarming queries can help. If you 
index frequently, it’s a balance between how much time you spend autowarming 
and how much time a user might have to wait.

In extreme cases, i.e. autoSoftCommits every second or 
hard-commit-with-opensearcher=true, it can be best to disable autowarming 
entirely.

And pretty often, I find just autowarming the queryResultCache and filterCache 
with relatively low numbers (< 20) works quite well.

I’ll also say that in the bad old days before docValues, autowarming had more 
of an impact if you faceted or sorted, as uninverting the structures on the 
heap could get very expensive. DocValues fields are much quicker to load, 
although they can still take a noticeable amount if time especially if there 
are lots…

FWIW,
Erick

> On Aug 8, 2020, at 9:48 AM, Dave  wrote:
> 
> Ah. Glad you found it. Yeah warming queries are much better substituted with 
> home made scripts if you need them. I like to use the previous days logs and 
> run the last couple hundred or so on a cron in the morning. 
> 
>> On Aug 8, 2020, at 9:39 AM, Schwartz, Tony  wrote:
>> 
>> I did not have a suggester set up.  I disabled the spell checker component, 
>> but that wasn't the problem.  I found my issue... it was related to a 
>> warming query i was running for each newly opened searcher.  Early on I 
>> enabled that, but I completely forgot about it.  And i don't believe it's 
>> needed.  I was hoping it would help with performance related to time 
>> filtering and sorting.  But, now it seems to be performing quite well 
>> without it.
>> 
>> Tony
>> 
>> 
>> 
>> From: Schwartz, Tony
>> Sent: Friday, August 7, 2020 6:27 PM
>> To: solr-user@lucene.apache.org
>> Subject: RE: solr startup
>> 
>> suggester?  what do i need to look for in the configs?
>> 
>> Tony
>> 
>> 
>> 
>> Sent from my Verizon, Samsung Galaxy smartphone
>> 
>> 
>> 
>>  Original message 
>> From: Dave 
>> mailto:hastings.recurs...@gmail.com>>
>> Date: 8/7/20 18:23 (GMT-05:00)
>> To: solr-user@lucene.apache.org
>> Subject: Re: solr startup
>> 
>> It sounds like you have suggester indexes being built on startup.  Without 
>> them they just come up in a second or so
>> 
>>> On Aug 7, 2020, at 6:03 PM, Schwartz, Tony 
>>> mailto:tony.schwa...@cinbell.com>> wrote:
>>> 
>>> I have many collections.  When I start solr, it takes 30 - 45 minutes to 
>>> start up and load all the collections.  My collections are named per day.  
>>> During startup, solr loads the collections in alpha-numeric name order.  I 
>>> would like solr to load the collections in the descending order.  So the 
>>> most recent collections are loaded first and are available for searching 
>>> while the older collections are not as important.  Is this possible?
>>> 
>>> 

Re: solr startup

[Dave]

Ah. Glad you found it. Yeah warming queries are much better substituted with 
home made scripts if you need them. I like to use the previous days logs and 
run the last couple hundred or so on a cron in the morning. 

> On Aug 8, 2020, at 9:39 AM, Schwartz, Tony  wrote:
> 
> I did not have a suggester set up.  I disabled the spell checker component, 
> but that wasn't the problem.  I found my issue... it was related to a warming 
> query i was running for each newly opened searcher.  Early on I enabled that, 
> but I completely forgot about it.  And i don't believe it's needed.  I was 
> hoping it would help with performance related to time filtering and sorting.  
> But, now it seems to be performing quite well without it.
> 
> Tony
> 
> 
> 
> From: Schwartz, Tony
> Sent: Friday, August 7, 2020 6:27 PM
> To: solr-user@lucene.apache.org
> Subject: RE: solr startup
> 
> suggester?  what do i need to look for in the configs?
> 
> Tony
> 
> 
> 
> Sent from my Verizon, Samsung Galaxy smartphone
> 
> 
> 
>  Original message 
> From: Dave mailto:hastings.recurs...@gmail.com>>
> Date: 8/7/20 18:23 (GMT-05:00)
> To: solr-user@lucene.apache.org
> Subject: Re: solr startup
> 
> It sounds like you have suggester indexes being built on startup.  Without 
> them they just come up in a second or so
> 
>> On Aug 7, 2020, at 6:03 PM, Schwartz, Tony 
>> mailto:tony.schwa...@cinbell.com>> wrote:
>> 
>> I have many collections.  When I start solr, it takes 30 - 45 minutes to 
>> start up and load all the collections.  My collections are named per day.  
>> During startup, solr loads the collections in alpha-numeric name order.  I 
>> would like solr to load the collections in the descending order.  So the 
>> most recent collections are loaded first and are available for searching 
>> while the older collections are not as important.  Is this possible?
>> 
>> 

RE: solr startup

[Schwartz, Tony]

I did not have a suggester set up.  I disabled the spell checker component, but 
that wasn't the problem.  I found my issue... it was related to a warming query 
i was running for each newly opened searcher.  Early on I enabled that, but I 
completely forgot about it.  And i don't believe it's needed.  I was hoping it 
would help with performance related to time filtering and sorting.  But, now it 
seems to be performing quite well without it.

Tony



From: Schwartz, Tony
Sent: Friday, August 7, 2020 6:27 PM
To: solr-user@lucene.apache.org
Subject: RE: solr startup

suggester?  what do i need to look for in the configs?

Tony



Sent from my Verizon, Samsung Galaxy smartphone



 Original message 
From: Dave mailto:hastings.recurs...@gmail.com>>
Date: 8/7/20 18:23 (GMT-05:00)
To: solr-user@lucene.apache.org
Subject: Re: solr startup

It sounds like you have suggester indexes being built on startup.  Without them 
they just come up in a second or so

> On Aug 7, 2020, at 6:03 PM, Schwartz, Tony 
> mailto:tony.schwa...@cinbell.com>> wrote:
>
> I have many collections.  When I start solr, it takes 30 - 45 minutes to 
> start up and load all the collections.  My collections are named per day.  
> During startup, solr loads the collections in alpha-numeric name order.  I 
> would like solr to load the collections in the descending order.  So the most 
> recent collections are loaded first and are available for searching while the 
> older collections are not as important.  Is this possible?
>
>