[
https://issues.apache.org/jira/browse/HBASE-19527?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16329729#comment-16329729
]
Appy commented on HBASE-19527:
------------------------------
Idk what's good - make them deamon thread or not. Been trying to reason in my
head for quite a while now, but can't answer for sure. Just adding some
thoughts here.
WorkerThread class in ProcExecutor can stop itself if it's idle and the server
is going down. But what about the threads doing actual work?
Making daemon thread
- Pro: master shutdown will not get stuck in any case
- Con: Our 'clean' shutdown isn't exactly clean, it'll crash end many
background work like create table, merge regions, and anything that we aim to
build on top of proc framework - backup, replication, etc
Not making threads daemon:
- Pro: Reliably ending ongoing work at defined sync points
- Cons: Master shutdown can get stuck
I personally hate landing in a place where our 'clean' shutdown will actually
be like a crash.
One alternative I have is:
Don't make them daemon threads. Start a ShutdownMonitor thread in
HMaster.stop() (which should be Daemon thread) and if it finds itself running
for more than X seconds, then call System.exit() (with a nice msg on why such
abruptness of course).
> Make ExecutorService threads daemon=true.
> -----------------------------------------
>
> Key: HBASE-19527
> URL: https://issues.apache.org/jira/browse/HBASE-19527
> Project: HBase
> Issue Type: Sub-task
> Reporter: stack
> Assignee: stack
> Priority: Major
> Fix For: 2.0.0-beta-2
>
> Attachments: HBASE-19527.branch-2.001.patch,
> HBASE-19527.branch-2.002.patch, HBASE-19527.master.001.patch,
> HBASE-19527.master.001.patch, HBASE-19527.master.001.patch,
> HBASE-19527.master.002.patch
>
>
> Let me try this. ExecutorService runs OPENs, CLOSE, etc. If Server is going
> down, no point in these threads sticking around (I think). Let me try this.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
