Re: Solr and Jetty and Servlets

Would be nicer if I could go through them and record a decision then get a green check mark so it was clear which ones were addressed vs punted but it also takes nearly forever to complete. On Sun, Nov 7, 2021 at 11:43 PM David Smiley wrote: > Sonatype-lift giving a passing review on a PR is not

Re: Solr and Jetty and Servlets

Sonatype-lift giving a passing review on a PR is not a qualification we have on our project to merge. Therefore it's also pointless to tell it to "ignore" anything. The bot provides feedback at specific line numbers that we should at least briefly consider; ideally say something about in the comm

Re: Solr and Jetty and Servlets

https://github.com/apache/solr/pull/397 is the new PR for those interested. Will try to merge it next weekend if no objections are found. Am a little unclear on what sonatype-lift is doing. I told it to ignore one (that I filed a separate ticket for) then made a fix thinking that that would cull o

Re: Solr and Jetty and Servlets

We do not currently have the docker image being built and tested on CI. I can get working on that. On Wed, Sep 1, 2021 at 3:07 PM David Smiley wrote: > JettySolrRunner can only do so much. There are other differences between > a real running server and JettySolrRunner such as our bin/solr scrip

Re: Solr and Jetty and Servlets

JettySolrRunner can only do so much. There are other differences between a real running server and JettySolrRunner such as our bin/solr script and probably much and all of our Jetty configuration. I've been bitten by a classpath bug that is really only possible to see when you run things for real

Re: Solr and Jetty and Servlets

Gus: I skimmed very little of this mail/thread -- just enough to recognize that it's a rabbit hole I'm not ready to devote brain cells to, but aplaud your interest in doing so. I will comment on just one aspect of your email, only to point you at some "prior hole diving" i did, on sub-topic y

Re: Solr and Jetty and Servlets

p just start Jetty > embedded, add listen ip, port numbers, etc. We can still use the config API > of Jetty for this, so maybe put this into an XML file (although XML is a > bad format for this stuff, goes back to web applications). Then Bootsrap > starts the Corecontainer and the embe

Re: Solr and Jetty and Servlets

uld be centralized. But this servlet would only take >> >>> care of the dispatching, so DispatchServlet would be fine. But it should >> >>> only dispatch, nothing else. >> >>> >> >>> My ideal Solr would be: Add some org.apache.

Re: Solr and Jetty and Servlets

): Maybe we should install a separate > ServletContextHandler/Classloader for each Core. This would allow us to go > away with our own classloader magic. Bootstrap could also do this: It > starts the CoreContainer, Zookeeper, Jetty Installs a servlet for the admin > UI and the Core/Col

Re: Solr and Jetty and Servlets

(not sure if they are already implemented in the patch): >>> Maybe we should install a separate ServletContextHandler/Classloader for >>> each Core. This would allow us to go away with our own classloader magic. >>> Bootstrap could also do this: It starts the C

Re: Solr and Jetty and Servlets

gt;> CoreContainer gets his own ServletContextHandler and can therefore easily >> be loaded unloaded from Jetty. >> >> I wrote many other apps in my daily live like this, the startup time >> (especially for microservices) is great. A Jetty starting up in >> milliseco

Re: Solr and Jetty and Servlets

> be loaded unloaded from Jetty. > > I wrote many other apps in my daily live like this, the startup time > (especially for microservices) is great. A Jetty starting up in > milliseconds and going to service is unbeatable. > > Uwe > > - > Uwe Schindler > Achterdiek

Re: Solr and Jetty and Servlets

etty starting up in milliseconds > and going to service is unbeatable. > > Uwe > > - > Uwe Schindler > Achterdiek 19, D-28357 Bremen > https://www.thetaphi.de <https://www.thetaphi.de/> > eMail: u...@thetaphi.de <mailto:u...@thetaphi.de> > > From:

RE: Solr and Jetty and Servlets

Miller Sent: Wednesday, August 18, 2021 9:08 AM To: Gus Heck Cc: dev@solr.apache.org Subject: Re: Solr and Jetty and Servlets Agreement or not, I wouldn’t step in front of whatever you did. I doubt anyone is going to take on actually pulling out of a web app because it’s easier to lean on it

Re: Solr and Jetty and Servlets

Agreement or not, I wouldn’t step in front of whatever you did. I doubt anyone is going to take on actually pulling out of a web app because it’s easier to lean on it for configuration and I doubt start up speed is a very high priority for most. It had never really occurred to me in the past that

Re: Solr and Jetty and Servlets

*ServletContextHandler not ServletContext On Tue, Aug 17, 2021 at 11:03 AM Gus Heck wrote: > Ok nope, since that only happens with WebAppContext, not ServletContext > (the parent class). whew :) > > On Tue, Aug 17, 2021 at 10:33 AM Gus Heck wrote: > >> WRT the JettySolrRunner tangent, I wonder

Re: Solr and Jetty and Servlets

Ok nope, since that only happens with WebAppContext, not ServletContext (the parent class). whew :) On Tue, Aug 17, 2021 at 10:33 AM Gus Heck wrote: > WRT the JettySolrRunner tangent, I wonder if our lack of a web.xml and > thus lack of metadata-complete=true in web.xml means that we wind up > s

Re: Solr and Jetty and Servlets

WRT the JettySolrRunner tangent, I wonder if our lack of a web.xml and thus lack of metadata-complete=true in web.xml means that we wind up scanning for annotations? On Tue, Aug 17, 2021 at 10:06 AM Gus Heck wrote: > https://issues.apache.org/jira/browse/SOLR-15590 > > On Tue, Aug 17, 2021 at 9:

Re: Solr and Jetty and Servlets

https://issues.apache.org/jira/browse/SOLR-15590 On Tue, Aug 17, 2021 at 9:44 AM Gus Heck wrote: > Ok well it would be interesting to compare quickstarting vs > JettySolrRunner, and reading up on quickstart > gives ideas about pre-building > the quickst

Re: Solr and Jetty and Servlets

Ok well it would be interesting to compare quickstarting vs JettySolrRunner, and reading up on quickstart gives ideas about pre-building the quickstart xml, but web.xml for JettySolrRunner is a tangent. Either way we are still in a container and I think I

Re: Solr and Jetty and Servlets

The downside to respecting web.xml and making JettySolrServer serve a webapp is that loading a webapp is very expensive and slow in comparison. JettySolrServer actually starts up extremely quickly. It’s almost more appealing to change the Server to use the JettySolrServer strategy. It’s so slow to

Re: Solr and Jetty and Servlets

So I'm not interested in *adhering* to anything here, just using stuff where it helps... I see tools sitting on the shelf, already bought and paid for and carried with us to every job but then they sit in the back of the truck mostly unused... and (I think) they look useful. This should not be see

Re: Solr and Jetty and Servlets

Sounds like an interesting adventure last weekend. I'm unclear what the point of going this direction is; my instinct is to go the opposite direction. You seem to suggest there are some simplification/organization benefits, which I love, so I'll need to look at what you've done to judge that for

Solr and Jetty and Servlets

*TLDR:* I've got a working branch where CoreContainer & our startup process is extracted from SolrDispatch Filter. Do other folks think that is interesting enough that I should make a JIRA and/or PR with intention to make this change? *Details:*