zookeeper-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alexander Shraer <shra...@gmail.com>
Subject Re: NewConfigNoQuorum?
Date Mon, 10 Nov 2014 22:34:27 GMT
Not sure. If you know how long state copying & syncing should take then you
could wait for that
time, retry and then throw it back if it still doesn't work.

On Mon, Nov 10, 2014 at 2:18 PM, Jordan Zimmerman <
jordan@jordanzimmerman.com> wrote:

> So, what’s the correct way to handle the exception? Should Curator throw
> it back to the user or retry?
>
> -JZ
>
>
> On November 10, 2014 at 5:14:26 PM, Alexander Shraer (shralex@gmail.com)
> wrote:
>
> This happens if you're trying to reconfigure to a new configuration S' but
> there is no quorum of S' connected
> and synced with the current leader. This could be either transient or more
> serious, depends...
> For example it could be the case that you are trying to add a few servers
> and there is a lot of state so
> syncing isn't complete yet and you just need to retry later. On the other
> hand it could be that S' includes
> servers that have crashed or don't exist, so retrying won't help.
>
> On Mon, Nov 10, 2014 at 8:54 AM, Ivan Kelly <ivank@apache.org> wrote:
>
> > Isn't this a transient error that occurs while the cluster is
> > reconfiguring? Seems to me that a retry would be correct.
> >
> > -Ivan
> >
> > On 10 November 2014 16:56, Jordan Zimmerman <jordan@jordanzimmerman.com>
>
> > wrote:
> > > In Alexander’s ZK meetup talk, he mentioned a new exception:
> > NewConfigNoQuorum. How should Curator handle this? Should the reconfig
> > operation just be retried when this is thrown or does it imply a deeper
> > problem?
> > >
> > > -Jordan
> > >
> > >
> >
>
>

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