Re: Clusterware and nodeid
From: Stefano Cislaghi <s.cislaghi_at_gmail.com>
Date: Fri, 24 Jun 2011 09:17:26 +0200
Message-ID: <BANLkTiktbW7o=sh7+N4g9rGAQe56e2KQPg_at_mail.gmail.com>
On 24 June 2011 09:12, D'Hooge Freek <Freek.DHooge_at_uptime.be> wrote:
> Martin,
>
> This would also mean that, when you do a dropnode procedure, this information is not removed.
>
> This could be done intentionally as part of the "grid plug and play" in which nodes can be "automatically" added / removed. Keeping the information would insure that a temporarily removed node keeps the same node id (and instance id) when added back to the cluster.
>
> It could also just be a bug ;-)
>
Date: Fri, 24 Jun 2011 09:17:26 +0200
Message-ID: <BANLkTiktbW7o=sh7+N4g9rGAQe56e2KQPg_at_mail.gmail.com>
On 24 June 2011 09:12, D'Hooge Freek <Freek.DHooge_at_uptime.be> wrote:
> Martin,
>
> This would also mean that, when you do a dropnode procedure, this information is not removed.
>
> This could be done intentionally as part of the "grid plug and play" in which nodes can be "automatically" added / removed. Keeping the information would insure that a temporarily removed node keeps the same node id (and instance id) when added back to the cluster.
>
> It could also just be a bug ;-)
>
Or just an 'happy feature' ... I'll try to open a SR to Oracle.
Stefano
-- http://www.freelists.org/webpage/oracle-lReceived on Fri Jun 24 2011 - 02:17:26 CDT