Re: solr and dovecot: high load

2022-12-01 Thread alessia . gagliano
Hello, thank you for the answers. We reduced the heap size to 20GB and it has made the situation better: it didn't stop, but the I/O activity was still high. We are a bit at a loss on what to monitor, maybe the garbage collector activity? I'm not aware of any particular planned recurrent

Wired behavior of maxClauseCount restriction since upgrading to solr 9.1

2022-12-01 Thread michael dürr
Hi, today we updated solr to version 9.1 (lucene version 9.3) Since then we noticed plenty of TooManyNestedClauses in the logs. Our setting for maxClauseCount is 1024 I played around a lot and could trace it down to this: * I built an index from scratch with two fields (id is unique key) and luce

Re: SOLR adding ,​ to strings erroneously

2022-12-01 Thread Eric Pugh
Shawn, Have we received a couple of mentions of this? Or am I misremembering? Do we need to open a JIRA and change how logging.js works? ERic > On Nov 30, 2022, at 11:44 PM, Shawn Heisey wrote: > > On 11/30/22 13:44, Matthew Castrigno wrote: >> Using SOLR 9.0 and the ScriptUpdatProcesor,

Question regarding Replica placement

2022-12-01 Thread Aurélien MAZOYER
Hello, I have some unexpected behavior regarding the replica placement. I am using solr 9.1.0 I start two nodes : solr.cmd start -c solr.cmd start -p 8984 -z localhost:9983 I add my replica policy: curl -X POST -H 'Content-type: application/json' -d '{ "add":{ "name": ".placeme

Re: Wired behavior of maxClauseCount restriction since upgrading to solr 9.1

2022-12-01 Thread Jan Høydahl
Have you tried using Terms Query? It is much more efficient than many boolean should clauses ?q={!terms f=id}1 2 3 4...1025 Jan > 1. des. 2022 kl. 13:27 skrev michael dürr : > > Hi, > > today we updated solr to version 9.1 (lucene version 9.3) > Since then we noticed plenty of TooManyNestedCl

RE: Very High CPU when indexing

2022-12-01 Thread Matias Laino
Hi Shawn, thanks again for the reply. I've tried increasing the memory to 32 gb and 16gb of ram heap with 8 cores, and even though I still see peaks of 300% CPU on the solr process it can handle it (solr doesn't go down). But, I've tried several different configurations for the auto commit and s

Re: SOLR adding ,​ to strings erroneously

2022-12-01 Thread dmitri maziuk
On 2022-12-01 6:41 AM, Eric Pugh wrote: Shawn, Have we received a couple of mentions of this? Or am I misremembering? Do we need to open a JIRA and change how logging.js works? https://issues.apache.org/jira/projects/SOLR/issues/SOLR-16469 It's interesting that the browser doesn't interpr

Re: SOLR adding ,​ to strings erroneously

2022-12-01 Thread Michael Conrad
We use the zero-width space trick here to ensure links break properly when formatting them for PDF and HTML display in our articles with extra long URLs. Inserting a regular space would work, but also would display incorrectly for human parsing. On 12/1/22 11:12, dmitri maziuk wrote: On 2022-

Solr Operator with a Custom Solr Image

2022-12-01 Thread Kira Traynor
Hi All, I want to use the Solr Operator (v0.6.0) with a custom solr image that is built from a forked version of solr. Would there be a way to replace the CRDs and use a different image tag in the solrcloud yaml? What would be the most straightforward way to do this? Thanks for the help! Kira

Re: SOLR adding ,​ to strings erroneously

2022-12-01 Thread Thomas Corthals
Op do 1 dec. 2022 om 17:12 schreef dmitri maziuk : > On 2022-12-01 6:41 AM, Eric Pugh wrote: > > Shawn, > > > > Have we received a couple of mentions of this? Or am I misremembering? > Do we need to open a JIRA and change how logging.js works? > > > https://issues.apache.org/jira/projects/SOLR/is

Re: Wired behavior of maxClauseCount restriction since upgrading to solr 9.1

2022-12-01 Thread Thomas Heigl
Hi Jan, We ran into the same issue. Terms queries sound like the ideal solution for our use case, but I couldn't find any documentation on the {!terms} syntax. Is there anything in the official docs? Best, Thomas On Thu, Dec 1, 2022 at 2:09 PM Jan Høydahl wrote: > Have you tried using Terms Q

Re: Wired behavior of maxClauseCount restriction since upgrading to solr 9.1

2022-12-01 Thread Kevin Risden
https://solr.apache.org/guide/solr/latest/query-guide/other-parsers.html#terms-query-parser The "!{terms ..." syntax is short for a query parser. Its a terms query parser and as Jan said its way more efficient than boolean clauses for a list of terms. Kevin Risden On Thu, Dec 1, 2022 at 1:04 PM

RE: Very High CPU when indexing

2022-12-01 Thread Matias Laino
I've tried with multiple different autosoft commit and auto commit configurations, and it always takes 2:30 - 3 minutes to get the records available on search, CPU is being pretty good since I upgraded, and memory should be plenty unless I'm mistaken, I'm lost at this point. Any help will be r

Re: Solr Operator with a Custom Solr Image

2022-12-01 Thread Jan Høydahl
Sure, see here https://github.com/apache/solr-operator/tree/main/docs/solr-cloud#official-solr-images Jan > 1. des. 2022 kl. 18:07 skrev Kira Traynor : > > Hi All, > > I want to use the Solr Operator (v0.6.0) with a custom solr image that is > built from a forked version of solr. Would there b

Re: Very High CPU when indexing

2022-12-01 Thread Jan Høydahl
What are your cache settings? Are you using autoWarmCount or explicit cache warming? It could be a source of long commit times. Jan > 1. des. 2022 kl. 22:35 skrev Matias Laino : > > > I've tried with multiple different autosoft commit and auto commit > configurations, and it always takes 2:30

Re: Wired behavior of maxClauseCount restriction since upgrading to solr 9.1

2022-12-01 Thread michael dürr
Thanks to all of you for your advice on using the terms query! I wasn't aware of this syntax until now. Anyways it would be good to know whether I hit a bug or not. Are my example queries probably rewritten to something that has more boolean clauses? If so, why doesn't that apply to the query for

Re: Wired behavior of maxClauseCount restriction since upgrading to solr 9.1

2022-12-01 Thread Jan Høydahl
A plain q=id:(a b c) is parsed into a boolean query with three SHOULD clauses, i.e. OR. Try to add &debugQuery=true to a request and see how it gets parsed. Then if the limit is 1024 you'll get errors above. Jan > 2. des. 2022 kl. 07:43 skrev michael dürr : > > Thanks to all of you for your ad