[jboss-jira] [JBoss JIRA] (WFLY-1870) Optimize read-resource-description() responses

[Heiko Braun (JIRA)]

     [ https://issues.jboss.org/browse/WFLY-1870?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Heiko Braun resolved WFLY-1870.
-------------------------------

        Labels:   (was: breaking_change)
    Resolution: Done


access-control=trim-descriptions does the job
                
> Optimize read-resource-description() responses
> ----------------------------------------------
>
>                 Key: WFLY-1870
>                 URL: https://issues.jboss.org/browse/WFLY-1870
>             Project: WildFly
>          Issue Type: Enhancement
>          Components: Domain Management
>            Reporter: Heiko Braun
>            Assignee: Brian Stansberry
>             Fix For: 8.0.0.CR1
>
>
> hbraun: kkhan: is there a wat to teach the wildcard handler for rrd not to include all resources?
> [5:47pm] hbraun: it's redundant information
> [5:47pm] hbraun: a way to teach
> [5:47pm] kkhan: in access-control, or in plain r-r-d?
> [5:47pm] kkhan: For plain r-r-d, I am not sure we can change that at this stage
> [5:47pm] hbraun: both shold be possible it hink
> [5:48pm] kkhan: since it has 'always' done it that way
> [5:48pm] hbraun: ok, and for access control?
> [5:49pm] kkhan: hbraun: I am not sure right now
> [5:49pm] hbraun: kkhan: so when you request /subsystem=datasources/datasource=*:rrd() you only get a single representation?
> [5:49pm] hbraun: kkhan: but do you see what I mean?
> [5:51pm] kkhan: hbraun: So access-control is only included in the datasource=>* entry and not in datasource=>ExampleDS?
> [5:51pm] bstansberry: sorry guys, somehow this room was closed in Colloquy
> [5:51pm] hbraun: kkhan: no, what i mean is:
> [5:52pm] hbraun: requests using wildcards to rrd() return a list of resources, including *
> [5:52pm] hbraun: but each specific one as well
> [5:52pm] hbraun: but the information returned is redundant
> [5:53pm] hbraun: i wam wondering if we can change it to be more optimized (not return a list, but a single * resource represenataiton)
> [5:53pm] hbraun: makes sense?
> [5:54pm] kkhan: hbraun: I have no idea why it was done that way 
> [5:55pm] hbraun: but tou see the problem now?
> [5:55pm] kkhan: But someone somewhere might be relying on that information
> [5:55pm] kkhan: or is this new behaviour?
> [5:55pm] hbraun: kkhan: that's no excuse not to optimize it some point
> [5:55pm] hbraun: no, it's always been like that
> [5:55pm] kkhan: I'll defer that one to bstansberry then 
> [5:55pm] hbraun: but the payload explodedes
> [5:56pm] hbraun: bstansberry: do you have a minute?
> [5:56pm] kkhan: I see what you mean though
> [5:56pm] hbraun: kkhan: great
> [5:56pm] kkhan: since the * and each subsystem will be the same static data
> [5:57pm] hbraun: kkhan: exactly
> [5:57pm] kkhan: and I guess each one gets access-control calculated as well
> [5:57pm] hbraun: and you've also provided the "exceptions" element for access control meta data
> [5:57pm] hbraun: +1
> [5:57pm] hbraun: yes
> [5:57pm] hbraun: a big mess
> [5:57pm] hbraun: cannot be low foot print
> [5:58pm] hbraun: it travels through the domain as well
> [5:59pm] hbraun: IMO this is one improvement with great impact on performance, but very little probelms for the clients
> [5:59pm] hbraun: besides: at the end of the day, we are the number one client anyway
> [6:00pm] hbraun: kkhan: anyway, i'll discuss it with bstansberry when he's available
> [6:01pm] hbraun: kkhan: but maybe that's something you can look at, while doing the other improvements?
> [6:01pm] hbraun: regarding the actual scope of work and implications?
> [6:01pm] kkhan: sure
> [6:01pm] kkhan: since every time I switch branches I seem to always come back to this 
> [6:02pm] hbraun: kkhan: i think it was just derived from the way read-resource works
> [6:03pm] hbraun: the result juggling in that case is correct
> [6:03pm] hbraun: just for rrd() it doesn't make sense
> [6:03pm] kkhan: hbraun: yeah I have the vague feeling too

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira