Re: [PR] SOLR-17525: Text Embedder Query Parser [solr]

2024-12-05 Thread via GitHub


alessandrobenedetti commented on PR #2809:
URL: https://github.com/apache/solr/pull/2809#issuecomment-2520803184

   Thanks guys, I updated the PR with additional minor changes in response to 
@dsmiley and other people feedback.
   
   I completely agree on merging this and iterate. It's not meant to be final 
in any way, to me it's just a big enabler, an entry point.
   
   It will be my pleasure to iterate on it when I have time and I encourage 
anyone interested to iterate on it and improve it!
   
   I'll leave it here another 3-4 days and then I'll merge it!
   In the meantime if you notice any blocker let me know!
   I'll also do all checks and gradle stuff to make sure it's ok


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



Re: [PR] SOLR-17525: Text Embedder Query Parser [solr]

2024-12-05 Thread via GitHub


alessandrobenedetti commented on code in PR #2809:
URL: https://github.com/apache/solr/pull/2809#discussion_r1871650626


##
solr/modules/llm/src/java/org/apache/solr/llm/texttovector/model/SolrTextToVectorModel.java:
##
@@ -63,12 +64,21 @@ public static SolrEmbeddingModel getInstance(
   var builder = modelClass.getMethod("builder").invoke(null);
   if (params != null) {
 /**

Review Comment:
   my bad, it's apparent I don't do comments very much, fixed!



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



Re: [PR] SOLR-17525: Text Embedder Query Parser [solr]

2024-12-05 Thread via GitHub


epugh commented on PR #2809:
URL: https://github.com/apache/solr/pull/2809#issuecomment-2520472327

   > Eric, you demonstrate how to save some arbitrary file into the cluster to 
the so-called "file store". But didn't demonstrate how anything in Solr would 
use it. But that's in a separate world/namespace from a new LLM-specific model 
store Alessandro wishes to add that is also distinct from LTR's similar one (do 
you know of it?). I'm dubious that we need ones specific to LLM or LTR. If the 
file store came first (it came after LTR actually), I suspect LTR's wouldn't 
exist and then Alessandro wouldn't have imitated it here. Both those by default 
save stuff in ZK in SolrCloud mode but are configurable to use the file system 
which could probably point to the file store. Oh my. Options, options; too many 
probably. Once added; good luck removing unless we are very clear about 
"experimental". Maybe Eric you have more familiarity and can add to the 
discussion.
   
   Honestly, I may be derailing this specific PR...   I would like us to move 
to a single more robust file store.  However, honestly, that is probably a 
orthagonal concern to this PR...   Let's get this in, and in the future, we can 
work on potentially consolidating down to one great solution.   I don't think 
this PR, using the Managed Resources more, makes it harder to eventually figure 
out how to retire/consolidate down the number of options!


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



Re: [PR] chore(deps): update org.apache.logging.log4j:* to v2.24.1 - autoclosed [solr]

2024-12-05 Thread via GitHub


risdenk commented on PR #2047:
URL: https://github.com/apache/solr/pull/2047#issuecomment-2520466098

   Ah thanks @janhoy not sure how I missed that comment. I'll take a closer look


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



[jira] [Updated] (SOLR-17585) CodeCache full during Gradle builds

2024-12-05 Thread Christos Malliaridis (Jira)


 [ 
https://issues.apache.org/jira/browse/SOLR-17585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Christos Malliaridis updated SOLR-17585:

Description: 
We are facing again issues similar to SOLR-16103 where we run out of CodeCache 
space. However, this may be caused in a different context now.

One assumption is that updating the CodeCache in our gradle.properties 
(template.gradle.properties) may fix the issue. Setting the jvm argument 
{{-XX:ReservedCodeCacheSize=120m}} may fix the issue, but it would be best to 
identify the cause, if possible.

One occurence of this issue is in 
[https://github.com/apache/solr/actions/runs/12073580646/job/33669959917]

(probably not the best job/commit to use for reproducing the issue).

  was:
We are facing again issues similar to SOLR-16103 where CodeCache becomes full. 
However, this may be caused in a different context now.

One assumption is that updating the CodeCache in our gradle.properties 
(template.gradle.properties) may fix the issue. Setting the jvm argument 
{{-XX:ReservedCodeCacheSize=120m}} may fix the issue, but it would be best to 
identify the cause, if possible.

One occurence of this issue is in 
[https://github.com/apache/solr/actions/runs/12073580646/job/33669959917]

(probably not the best job/commit to use for reproducing the issue).


> CodeCache full during Gradle builds
> ---
>
> Key: SOLR-17585
> URL: https://issues.apache.org/jira/browse/SOLR-17585
> Project: Solr
>  Issue Type: Improvement
>  Security Level: Public(Default Security Level. Issues are Public) 
>  Components: Gradle
>Reporter: Christos Malliaridis
>Priority: Major
>
> We are facing again issues similar to SOLR-16103 where we run out of 
> CodeCache space. However, this may be caused in a different context now.
> One assumption is that updating the CodeCache in our gradle.properties 
> (template.gradle.properties) may fix the issue. Setting the jvm argument 
> {{-XX:ReservedCodeCacheSize=120m}} may fix the issue, but it would be best to 
> identify the cause, if possible.
> One occurence of this issue is in 
> [https://github.com/apache/solr/actions/runs/12073580646/job/33669959917]
> (probably not the best job/commit to use for reproducing the issue).



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



Re: [PR] SOLR-16781: Remove solrconfig.xml directives [solr]

2024-12-05 Thread via GitHub


dsmiley commented on PR #2875:
URL: https://github.com/apache/solr/pull/2875#issuecomment-2519697587

   > My vote would be to tear it out. 
   
   +1
   I'm glad to hear you suggest this.  I thought I might have been alone in 
this thinking.  I don't think we should "support" this idea; it's too 
burdensome for us to uphold this concept to a good standard.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



Re: [PR] Solr 12276 angularjs to angular migration [solr]

2024-12-05 Thread via GitHub


NikolaiJDev commented on PR #2818:
URL: https://github.com/apache/solr/pull/2818#issuecomment-2520090026

   
   > Yes, what I (and probably others as well) would like to see is how it 
would integrate into the current project (which is as described above an open 
TODO), how someone would work and test the code base (like running the UI with 
`npm ng serve`), and where the webapp can be accessed / seen (like at 
`localhost:3000`). I believe the module's / angular project's README file is 
the best place to tell other developers how to get started with this 
sub-project. Angular already provides some basic information, this could be 
extended.
   
   
![image](https://github.com/user-attachments/assets/0bd2c41a-4c05-4be8-be10-c9c917a197cc)
   It will be integrated same way as it was done before because it has to be 
shipped together with main application then using Gradle service is a great 
practice, you asked me to be waved out current application from new one, so I 
removed the concept of integration for time being. However, the refactoring of 
the Gradle job and few changes to the web.xml might be required to provide to 
fully support the current approach. As I've already mentioned to you previously 
- It might be also required to make some tweaks and changes in the Backend 
filter, what we can only understand on the stage of integration, I hope it 
won't be required, but It may.
   
   > 
   > I meant starting and poking around, yes. So there is no state yet that 
makes request to a Solr backend, is that correct?
   > 
   > I think you removed a few too manyfiles in your latest updates. The 
environment files under `/src/environments` are missing and causing an error 
when launching the app with `ng serve`.
   > 
   > And one last tip, it's fine and also recommended to include 
`package-lock.json` in the VCS, as it ensures reproducable states of the 
project. :)
   
   Well, git shows me that envs are in place, I trigger the update as well for 
that.
   As well, I want to inform to run correctly app, you need to have either 
docker container with working solr application or it must run from the same 
hostname, just for clarification localhost that you are running from isn't same 
as for backend, CORS will appear to avoid it you need to make a proxy 
configuration such as Nginx server which will locate UI and middle tier 
application under same umbrella scope. It will help browser to avoid this 
issue. Because we will emulate same patter as if it will be run from same space 
or machine as backend application.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



Re: [PR] chore(deps): update org.apache.logging.log4j:* to v2.24.1 - autoclosed [solr]

2024-12-05 Thread via GitHub


janhoy commented on PR #2047:
URL: https://github.com/apache/solr/pull/2047#issuecomment-2520020161

   > @janhoy where did you see that log4j requires JDK 17? I don't see that 
based on https://logging.apache.org/log4j/2.x/release-notes.html
   > 
   > I found 
[apache/logging-log4j2#1922](https://github.com/apache/logging-log4j2/issues/1922)
 that affects the 2.21 version we use in Solr and we use the `notEmpty` so it 
actually produces more verbose logging.
   > 
   > I think we can upgrade log4j2
   
   If you can make it work without compile failure then go for it. Last I tried 
I ran into issues 
https://github.com/apache/solr/pull/2047#issuecomment-1971765269


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



[jira] [Created] (SOLR-17585) CodeCache full during Gradle builds

2024-12-05 Thread Christos Malliaridis (Jira)
Christos Malliaridis created SOLR-17585:
---

 Summary: CodeCache full during Gradle builds
 Key: SOLR-17585
 URL: https://issues.apache.org/jira/browse/SOLR-17585
 Project: Solr
  Issue Type: Improvement
  Security Level: Public (Default Security Level. Issues are Public)
  Components: Gradle
Reporter: Christos Malliaridis


We are facing again issues similar to SOLR-16103 where CodeCache becomes full. 
However, this may be caused in a different context now.

One assumption is that updating the CodeCache in our gradle.properties 
(template.gradle.properties) may fix the issue. Setting the jvm argument 
{{-XX:ReservedCodeCacheSize=120m}} may fix the issue, but it would be best to 
identify the cause, if possible.

One occurence of this issue is in 
[https://github.com/apache/solr/actions/runs/12073580646/job/33669959917]

(probably not the best job/commit to use for reproducing the issue).



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org



Re: [I] [Regression] security.json is not uploaded during the first initialization of SolrCloud [solr-operator]

2024-12-05 Thread via GitHub


mcarroll1 commented on issue #720:
URL: https://github.com/apache/solr-operator/issues/720#issuecomment-2521532629

   This is also blocking an upgrade for our team. Per @erwanval , it does seem 
like the zkcli returns error code `1` regardless of the issue. Would it be 
acceptable to rely on the string error messages (e.g. `NoNode for 
/security.json`)?  I'm interested in working on this next week if that solution 
is acceptable, or if we can find a reasonable alternative.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


-
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org