You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by Stack <st...@duboce.net> on 2013/06/24 22:49:16 UTC
Opinions on the 'watcher' script that surrounds launched hbase daemons
New in 0.95, when you do a process listing, there are two 'hbase'
processes. There is the old java process itself and then there is the
launcher script that is just hanging around waiting on the java process to
exit so it can clear the process's znode. Clearing the znode will mean the
master will know sooner of the missing regionserver. Otherwise it would
have to wait wait on the melting of the ephemeral znode (tens of seconds).
This 'watcher' was added by:
HBASE-5844 Delete the region servers znode after a regions server crash
What do folks think about how this facility has been implemented. I am not
questioning the merit of fast-clear of the znode. This is a good idea. I
just want to talk about how it is done. I am not a fan [1]. Our little
wrapper script verges into the process supervision space [2]. Rather than
do up our own, instead should we ship templates for a few common
supervisors and undo our custom one?
Thoughts?
St.Ack
1.
https://issues.apache.org/jira/browse/HBASE-5844?focusedCommentId=13463409&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13463409
2. http://en.wikipedia.org/wiki/Process_supervision
Re: Opinions on the 'watcher' script that surrounds launched hbase daemons
Posted by Jean-Daniel Cryans <jd...@apache.org>.
Yes.
J-D
On Mon, Jun 24, 2013 at 1:49 PM, Stack <st...@duboce.net> wrote:
> New in 0.95, when you do a process listing, there are two 'hbase'
> processes. There is the old java process itself and then there is the
> launcher script that is just hanging around waiting on the java process to
> exit so it can clear the process's znode. Clearing the znode will mean the
> master will know sooner of the missing regionserver. Otherwise it would
> have to wait wait on the melting of the ephemeral znode (tens of seconds).
>
> This 'watcher' was added by:
>
> HBASE-5844 Delete the region servers znode after a regions server crash
>
> What do folks think about how this facility has been implemented. I am not
> questioning the merit of fast-clear of the znode. This is a good idea. I
> just want to talk about how it is done. I am not a fan [1]. Our little
> wrapper script verges into the process supervision space [2]. Rather than
> do up our own, instead should we ship templates for a few common
> supervisors and undo our custom one?
>
> Thoughts?
> St.Ack
>
>
> 1.
> https://issues.apache.org/jira/browse/HBASE-5844?focusedCommentId=13463409&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13463409
> 2. http://en.wikipedia.org/wiki/Process_supervision
Re: Opinions on the 'watcher' script that surrounds launched hbase daemons
Posted by Stack <st...@duboce.net>.
Our Gregory did it a while back: HBASE-7386 Let me pull it into 0.95....
Thanks all,
St.Ack
On Thu, Jun 27, 2013 at 3:37 PM, Dave Wang <ds...@cloudera.com> wrote:
> +1 as well. Should we open a new JIRA to track this work?
>
> - Dave
>
>
> On Tue, Jun 25, 2013 at 11:18 AM, Enis Söztutar <en...@gmail.com>
> wrote:
>
> > +1 on not doing our own.
> >
> > Related is https://issues.apache.org/jira/browse/HBASE-7334
> >
> > Enis
> >
> >
> > On Mon, Jun 24, 2013 at 6:59 PM, Andrew Purtell <ap...@apache.org>
> > wrote:
> >
> > > On Mon, Jun 24, 2013 at 4:13 PM, Nick Dimiduk <nd...@gmail.com>
> > wrote:
> > >
> > > > It sounds like it's encroaching on BigTop territory.
> > > >
> > >
> > > I agree. Can we handle this there? Point a section of the book on
> > watchers
> > > to it?
> > >
> > >
> > > --
> > > Best regards,
> > >
> > > - Andy
> > >
> > > Problems worthy of attack prove their worth by hitting back. - Piet
> Hein
> > > (via Tom White)
> > >
> >
>
Re: Opinions on the 'watcher' script that surrounds launched hbase daemons
Posted by Dave Wang <ds...@cloudera.com>.
+1 as well. Should we open a new JIRA to track this work?
- Dave
On Tue, Jun 25, 2013 at 11:18 AM, Enis Söztutar <en...@gmail.com> wrote:
> +1 on not doing our own.
>
> Related is https://issues.apache.org/jira/browse/HBASE-7334
>
> Enis
>
>
> On Mon, Jun 24, 2013 at 6:59 PM, Andrew Purtell <ap...@apache.org>
> wrote:
>
> > On Mon, Jun 24, 2013 at 4:13 PM, Nick Dimiduk <nd...@gmail.com>
> wrote:
> >
> > > It sounds like it's encroaching on BigTop territory.
> > >
> >
> > I agree. Can we handle this there? Point a section of the book on
> watchers
> > to it?
> >
> >
> > --
> > Best regards,
> >
> > - Andy
> >
> > Problems worthy of attack prove their worth by hitting back. - Piet Hein
> > (via Tom White)
> >
>
Re: Opinions on the 'watcher' script that surrounds launched hbase daemons
Posted by Enis Söztutar <en...@gmail.com>.
+1 on not doing our own.
Related is https://issues.apache.org/jira/browse/HBASE-7334
Enis
On Mon, Jun 24, 2013 at 6:59 PM, Andrew Purtell <ap...@apache.org> wrote:
> On Mon, Jun 24, 2013 at 4:13 PM, Nick Dimiduk <nd...@gmail.com> wrote:
>
> > It sounds like it's encroaching on BigTop territory.
> >
>
> I agree. Can we handle this there? Point a section of the book on watchers
> to it?
>
>
> --
> Best regards,
>
> - Andy
>
> Problems worthy of attack prove their worth by hitting back. - Piet Hein
> (via Tom White)
>
Re: Opinions on the 'watcher' script that surrounds launched hbase daemons
Posted by Andrew Purtell <ap...@apache.org>.
On Mon, Jun 24, 2013 at 4:13 PM, Nick Dimiduk <nd...@gmail.com> wrote:
> It sounds like it's encroaching on BigTop territory.
>
I agree. Can we handle this there? Point a section of the book on watchers
to it?
--
Best regards,
- Andy
Problems worthy of attack prove their worth by hitting back. - Piet Hein
(via Tom White)
Re: Opinions on the 'watcher' script that surrounds launched hbase daemons
Posted by Nick Dimiduk <nd...@gmail.com>.
The template approach is cleaner. It sounds like it's encroaching on BigTop
territory.
On Mon, Jun 24, 2013 at 1:49 PM, Stack <st...@duboce.net> wrote:
> New in 0.95, when you do a process listing, there are two 'hbase'
> processes. There is the old java process itself and then there is the
> launcher script that is just hanging around waiting on the java process to
> exit so it can clear the process's znode. Clearing the znode will mean the
> master will know sooner of the missing regionserver. Otherwise it would
> have to wait wait on the melting of the ephemeral znode (tens of seconds).
>
> This 'watcher' was added by:
>
> HBASE-5844 Delete the region servers znode after a regions server crash
>
> What do folks think about how this facility has been implemented. I am not
> questioning the merit of fast-clear of the znode. This is a good idea. I
> just want to talk about how it is done. I am not a fan [1]. Our little
> wrapper script verges into the process supervision space [2]. Rather than
> do up our own, instead should we ship templates for a few common
> supervisors and undo our custom one?
>
> Thoughts?
> St.Ack
>
>
> 1.
>
> https://issues.apache.org/jira/browse/HBASE-5844?focusedCommentId=13463409&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13463409
> 2. http://en.wikipedia.org/wiki/Process_supervision
>