Re: A redundent two-server setup?

[Alex Davies]

Thank-you so much. That was it!

I am going to write an equivalent to the "MySQL Manual | 16.3.3 Quick
Test Setup of MySQL Cluster" but for the two servers and a management
server which I hope will be useful to the large number of mysql users
who are dealing with two webservers behind a loadbalancer/dns round
robin.

With many thanks,

Alex


On Tue, 28 Dec 2004 20:07:24 +0100, Mikael Ronström <mikael@stripped> wrote:
> Hi Alex,
> When you start after changing NoOfReplicas do you use --initial when
> starting the ndbd nodes.
> It is not possible to change this parameter once the initial start have
> been performed.
> 
> Rgrds Mikael
> 
> 2004-12-28 kl. 19.25 skrev Alex Davies:
> 
> > Dear Michael,
> >
> > If you want access to the servers please let me know.
> >
> > I have now repeated the process of trying it with NoOfReplicas=1 (and
> > it works). I then change NoOfReplicas=2, stop everything (checking
> > with ps aux | grep ndb) and then start it all. It does not work. I
> > then change NoOfReplicas=1, stop it all and start it and it all works.
> > So it is certainly that!
> >
> > If I start ndbd on each server using --nostart, and then on the mgm
> > server start one server (using id START) it "starts". after about 5
> > minutes it chages to "stage 2" and then hangs. As soon as I start the
> > other node both die straight away.
> >
> > With many thanks,
> >
> > Alex
> >
> >
> > On Tue, 28 Dec 2004 19:21:23 +0100, Mikael Ronström
> <mikael@stripped>
> > wrote:
> >> Hi Alex,
> >>
> >> 2004-12-28 kl. 18.58 skrev Alex Davies:
> >>
> >>> Thanks for your help. I must confess that although I am usually a
> >>> competent systems administrator I am not a C programmer so slightly
> >>> lost you in your previous email!
> >>>
> >>
> >> No problem :), the info you supplied was the correct
> >> I'll see what can be made out of it. I can see what went wrong
> >> but fail to understand why on a short look at it.
> >>
> >>> I think that the bit of the trace you are interested is this:
> >>>
> >>> NDBFS   000915 000917 000728 000934
> >>> CMVMI   000450 000458
> >>> QMGR    001815 001833
> >>>
> >>> --------------- Signal ----------------
> >>> r.bn: 252 "QMGR", r.proc: 3, r.sigId: 646 gsn: 235 "DISCONNECT_REP"
> >>> prio: 0
> >>> s.bn: 254 "CMVMI", s.proc: 3, s.sigId: 645 length: 2 trace: 8 #sec: 0
> >>> fragInf: 0
> >>>
> >>
> >>> Please let me know if you want any other information. Can I just
> >>> confirm that it *is* possible to run a totally redundent cluster on
> >>> three servers - two API/storage nodes and one node for managment.
> >>> (when I say totally redundent I mean you can pull the power plug out
> >>> of any one server with no effect on the overall system).
> >>>
> >>
> >> Yes, this is possible. If you pull the plug out of the storage/API
> >> nodes then there will be a few
> >> aborted transactions but then it will continue to work and these nodes
> >> can rejoin the cluster
> >> while in operation. Losing the management server has no impact on the
> >> system other than that
> >> nodes are not able to start if they were down and also there will be
> >> no
> >> cluster log while it is down.
> >> But the cluster will continue to process transactions even when the
> >> management server is down
> >> or one of the storage/API nodes.
> >>
> >> Rgrds Mikael
> >>
> >>> With many thanks,
> >>>
> >>> Alex
> >>>
> >>> --
> >>> MySQL Cluster Mailing List
> >>> For list archives: http://lists.mysql.com/cluster
> >>> To unsubscribe:
> >>> http://lists.mysql.com/cluster?unsub=1
> >>>
> >>>
> >> Mikael Ronström, Senior Software Architect
> >> MySQL AB, www.mysql.com
> >>
> >> Clustering:
> >> http://www.infoworld.com/article/04/04/14/HNmysqlcluster_1.html
> >>
> >> http://www.eweek.com/article2/0,1759,1567546,00.asp
> >>
> >>
> >
> >
> > --
> > Alex Davies // http://www.davz.net
> >
> > This email and any files transmitted with it are confidential and
> > intended solely for the use of the individual or entity to whom they
> > are addressed. If you have received this email in error please notify
> > the sender immediately by e-mail and delete this e-mail permanently.
> >
> Mikael Ronström, Senior Software Architect
> MySQL AB, www.mysql.com
> 
> Clustering:
> http://www.infoworld.com/article/04/04/14/HNmysqlcluster_1.html
> 
> http://www.eweek.com/article2/0,1759,1567546,00.asp
> 
> 


-- 
Alex Davies // http://www.davz.net

This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the sender immediately by e-mail and delete this e-mail permanently.