Dave Marion

Results 422 comments of Dave Marion

I kicked off a full build with ITs and the only ITs that failed are the same ones that are currently failing in `main`.

> They [scan hints] were never intended to change anything about data returned I equate scan hints with Oracle's [optimizer hints](https://docs.oracle.com/cd/B19306_01/server.102/b14211/hintsref.htm#i17496) which allow the user to alter the execution plan...

> I kinda like ScanServerChooser I'm partial to `ScanServerSelector`.

@keith-turner - I think your latest commit left `ScanServerIT.testBatchScannerTimeout` in a broken state.

@ctubbsii - I believe that I have addressed most, if not all, of your comments either by responding, creating a follow-on issue, or fixing the code. @ctubbsii @keith-turner @milleruntime @EdColeman...

Looks like you had planned on adding tests at some point. They are no longer needed?

> I really think the logic of this class should exist in the TabletServerResourceManager The JVM GC Logger is used by all server processes, not just the TabletServer.

I'm not sure that these warning indicate an error *will* occur based on [this](https://bugs.openjdk.org/browse/JDK-8160675?focusedCommentId=14173534&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-14173534) comment and this comment in https://bugs.openjdk.org/browse/JDK-8274336: > It is not necessarily erroneous that a field's type...

> If we can fix the warnings without touching the API like I did in https://github.com/apache/accumulo/pull/2777, then I think that would be good to get a head start. Yeah, I'm...

User was running tests against 2.1.0-SNAPSHOT and found that getting properties was consuming an unexpected amount of CPU time. User provided stack trace: ``` "Split/MajC initiator" #51 daemon prio=5 os_prio=0...