zookeeper-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Pramod Biligiri <pramodbilig...@gmail.com>
Subject Re: Partitioned Zookeeper
Date Mon, 19 May 2014 05:38:20 GMT
Hi Ted,
I see your point. You are right that it could be tricky.

Do you see any other problems with the approach I'm taking. If we see any
gains from it, we can look at the tricky issues next.


On Sun, May 18, 2014 at 9:40 PM, Ted Dunning <ted.dunning@gmail.com> wrote:

> On Sun, May 18, 2014 at 9:36 PM, Pramod Biligiri
> <pramodbiligiri@gmail.com>wrote:
> > I guess you mean that you can't parallellize the workload because a multi
> > command might require locking all the containers? Let me know if I'm
> > missing something.
> >
> Right.  Getting that to work cleanly could be difficult
> The problem is that once a transaction is accepted by one partition, it can
> be very difficult to unwind.  That means that if part of a multi is on one
> partition and another part is on another partition, you can't really do a
> 2-phase commit sort of thing since you can't roll back the part that worked
> when you find out about the part that didn't.

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message