Skip to content

Instantly share code, notes, and snippets.

@AdamBien
Created November 6, 2018 06:22
Show Gist options
  • Save AdamBien/d6af498fafa2c1dc8497f254140e6a31 to your computer and use it in GitHub Desktop.
Save AdamBien/d6af498fafa2c1dc8497f254140e6a31 to your computer and use it in GitHub Desktop.
57thAirhacksQ&A.md

Ask questions and see you at December, 3rd, 6.PM. CET: http://www.ustream.tv/channel/adambien

Also checkout recent episode:

56th airhacks.tv

Please keep the questions Java EE-stic. Means: as short and as concise as only possible.

@suchwerk
Copy link

suchwerk commented Dec 1, 2018

Hi Adam, some of my microservices depend on apache poi which leads to a war >18MB. How in general do you build an application server (docker) including static dependencies (which could also easily be updated)? Thank you from Landshut :)

@dempile
Copy link

dempile commented Dec 1, 2018

Hi Adam

  • After January 2019 we will have to pay for Oracle JDK 8, I see that Arun Gupta and James Gosling are launching Amazon Corretto with a long term support to 2023 https://www.youtube.com/watch?v=SLi6_ec29ek.. Have you use OpenJDK in production environment.
    Thanks.

@whyvrafvr
Copy link

Hi Adam. I was reading your post about ManagedScheduledExecutorService replacing EJB Timer. Do you have an idea how to mock this resource using Weld (I've read something about binding a resource building the WeldInitiator but it seems weird...)

@Vagspath
Copy link

Vagspath commented Dec 3, 2018

Hi Adam,

As i can understand @Inject does not use pooling vs @ejb which uses pooling. This means that it is slower if we use @Inject due to object creation?

@istovatis
Copy link

Hi Adam,
as consistency of data across microservices is not handled by the relational databases any more,
are you favour of manually handling the consistency (by adopting two-phase commit, retries logic etc) or adopting
a tool that provides eventual consistency, such as an event sourcing tool, for example debezium?

Thanks

@jessefarinacci
Copy link

Hi Adam,

Have you made any example attempts to configure your Docklands projects with microprofile-friendly writing of all outputs to stdout? There seems to be a diverse set of behaviors across application servers for what to put to initial stdout, when to cut over to a log file, where to write error messages, and what general log files to produce; let alone roll over frequency, etc etc etc. It would be nice, and best practice, to just dump all logging to stdout while in a container.

Bonus question if you have time: do logging frameworks have any future given increased containerization efforts?

Thanks!

@vkalayda
Copy link

vkalayda commented Dec 20, 2018

Hi Adam,

As i can understand @Inject does not use pooling vs @ejb which uses pooling. This means that it is slower if we use @Inject due to object creation?

Hello, as Adam says it depends on Application Server, as i know for instance WildFly does not use pools.

Regarding behaviour on GlassFish / Payara Application Servers:

If you are injecting component via @Inject and this is Stateless component - it will use object pools.
If you are injecting POJO with @Inject - it depends on scope of injected component.

So my recommendation is - use @Inject everywhere.

Thanks.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment