Martin Kouba

> I'm thinking something that does everything the initWeld() method you linked does. Something that can register the rest of the beans for you rather than relying on auto-discovery (like...

Hm, so you're proposing some kind of semi-automatic discovery ;-). > A CDI extension could work, probably using BeforeBeanDiscovery, ... This could work but I'm not sure it's worth a...

Apparently it's not a Quarkus-specific issue - see for example https://github.com/quartz-scheduler/quartz/issues/669. The suggested solution is to use a different SQL select for MS SQL server: `SELECT * FROM {0}LOCKS UPDLOCK...

> Do you have any idea why this issue started only after Quarkus 3.2.x? It was working fine until Quarkus 3.1.x I have no idea. The quartz version is the...

@chvsnraju I can see the problem now. There was an issue introduced in https://github.com/quarkusio/quarkus/pull/33534 (3.2.0.CR1) and it was fixed in https://github.com/quarkusio/quarkus/pull/35255 (3.2.4.Final). So an upgrade to Quarkus `3.2.4.Final` should solve...

I see. Well, then I'm running out of ideas... I suppose that you have the `quarkus.datasource.db-kind=mssql` config property in your project? Could you share the whole quartz config section from...

Hm, your `application.properties` is missing the `quarkus.quartz.clustered=true` and it seems that `quarkus.quartz.select-with-lock-sql` is [ignored unless the clustered mode is enabled](https://github.com/quarkusio/quarkus/blob/main/extensions/quartz/runtime/src/main/java/io/quarkus/quartz/runtime/QuartzSchedulerImpl.java#L594-L602)...

> @mkouba - Do you want me to close this, or will this be addressed in future releases without additional SQL as property? Well, I'm not quite sure if this...

So I ran the microbenchmarks and the only benchmark that shows _significant_ improvement is the [`RequestContextActivationBenchmark`](https://github.com/mkouba/arc-benchmarks/blob/master/src/main/java/io/quarkus/arc/benchmarks/RequestContextActivationBenchmark.java) which merely tests request context activation/deactivation, no bean operations - and it's expected due...

> Probably [RequestContextBenchmark](https://github.com/mkouba/arc-benchmarks/blob/master/src/main/java/io/quarkus/arc/benchmarks/RequestContextBenchmark.java), could have been relevant if the benchmark had a parameter to decide how many of these beans are computed eg for a total of 5 beans, just...