From: Date: March 8 2010 12:23pm Subject: RE: about how to implement load balancing and fail over automatically List-Archive: http://lists.mysql.org/cluster/7510 Message-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="gb2312" Content-Transfer-Encoding: quoted-printable Hi Ulherr, Thanks for your answer. You means technically it is there already for Solaris Cluster to = implement load balance for MySQL Cluster. But in fact no project has = used it as a load balancer formally? Do you think it is the best choice to use Solaris Cluster as a load = balancer between the application and MySQL Cluster? We know, for Linux operating system, there is a LVS. In your opinion, = can the Solaris Cluster replace LVS in the platform of Solaris? Thanks, Peter -----Original Message----- From: Detlef.Ulherr@stripped [mailto:Detlef.Ulherr@stripped]=20 Sent: 2010=C4=EA3=D4=C28=C8=D5 19:10 To: Yousuf Cc: Peter Chen; Andrew.Morgan@stripped; cluster@stripped; Burt = Clouse Subject: Re: about how to implement load balancing and fail over = automatically Hi Yousuf et all, I have good news, partially at least. If you combine Oracle Solaris Cluster (AKA Sun Cluster) and MySQL Cluster it is doable. Sun Cluster offers the concept of scalable services which is perfect for loadbalanced sql nodes. At least as long as ip based loadbalancing is enough. All the necessary changes are already in the SC product and QA is done.We are just lacking docs to announce the support for the combination of MySQL Cluster and Oracle Solaris Cluster. In terms of the client configuration, you then will just configure our scalable address and OSC is handling the loadbalancing. So technically it is there already, we just need a business case to put more weight on the docs priority, to make it available for everybody( cc'ed Burt). As soon as it is documented and announced, just go ahead. If it is for a proof of concept I can make a cheat sheet (howto) = available. Detlef Yousuf wrote: > On Solaris platform, you could look for Solaris cluster to provide HA > of SQL server in Active-Active mode and also use the Solaris cluster > to distribute the load among the 2 HA SQL servers. > For more details on the same, probably Detlef could share his > thoughts, I have copied him on this thread. > > regards > Yousuf > > Peter Chen wrote: >> Hi all, >> The best way to do this(to implement load balancing and fail over = automatically) is to use a software loadbalancer like Ldirectord, = coupled with a failover mechanism such as heartbeat. >> >> My Operating system is Solaris 10 sparc but not Linux. So there no = LVS on it. >> >> I want to know if there is a counterpart software like LVS on Solaris = 10 operating system? >> >> If not, I think I can't use this solution. >> Expect your answer, thanks. >> Peter >> >> >> >> -----Original Message----- >> From: Richard McCluskey [mailto: Richard McCluskey]=20 >> Sent: 2010=C4=EA1=D4=C213 >> To: Peter Chen >> Cc: cluster@stripped >> Subject: RE: about how to implement load balancing and fail over = automatically >> I second that, we use it to load balance the two SQL nodes in our >> mysql-cluster, and in the last 42 hours it has handled : >> >> >> Questions: 78,708,953 Slow queries: 102 Opens: 1,487,485 Flush >> tables: 2 Open tables: 64 Queries per second avg: 508.850 >> >> >> >> We have also watched it 'failover' recently when one of the SQL nodes >> wedged (faulty memory). The failover happened seamlessly, and without >> interrupting service. >> -----Original Message----- >> From: Andrew.Morgan@stripped [mailto:Andrew.Morgan@stripped]=20 >> Sent: 2010=C4=EA1=D4=C212=C8=D5 16:54 >> To: Peter Chen >> Cc: cluster@stripped >> Subject: RE: about how to implement load balancing and fail over = automatically >> >> Hi Peter, >> >> I've yet to try setting it up myself but if you're using MySQL's = JDBC >> connector then it can handle the failover for you. >> >> Regards, Andrew. >> >> =20 >>> -----Original Message----- >>> From: Peter Chen [mailto:peter.chen@stripped] >>> Sent: 12 January 2010 08:50 >>> To: cluster@stripped >>> Subject: about how to implement load balancing and fail over >>> automatically >>> >>> >>> >>> Now in my software system, I used MySQL as the backend DB server. >>> Because of the performance problem, I want to upgrade to MySQL = Cluster. >>> >>> I want to use MySQL Cluster NDB 7.0 >>> = , >>> I >>> have read this article: A.10. MySQL 5.1 FAQ - MySQL Cluster >>> >> cluster.html#qandaite >>> m-23-10-1-17> , and I know this: >>> >>> >>> >>> 23.10.14: How do I continue to send queries in the event that one of >>> the >>> SQL nodes fails? >>> >>> MySQL Cluster does not provide any sort of automatic failover = between >>> SQL nodes. Your application must be prepared to handle the loss of = SQL >>> nodes and to fail over between them. >>> >>> >>> >>> My question is that I don't know how to modify my code to implement >>> load >>> balancing and handle the loss of SQL nodes and to fail over between >>> them. >>> >>> Just modify the configuration of JDBC or use some third-party = software >>> to implement load balancing and fail over automatically? >>> >>> Expect your help, thanks. >>> >>> >>> =20 >> >> >> >> -- >> MySQL Cluster Mailing List >> For list archives: http://lists.mysql.com/cluster >> To unsubscribe: = http://lists.mysql.com/cluster?unsub=3Dmohammed.yousuf@stripped >> >> =20 --=20 *************************************************************************= **** Detlef Ulherr Staff Engineer Tel: (++49 6103) 752-248 Availability Engineering Fax: (++49 6103) 752-167 Sun Microsystems GmbH =20 Amperestr. 6 mailto:detlef.ulherr@stripped 63225 Langen http://www.sun.de/ *************************************************************************= **** Sitz der Gesellschaft: Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten Amtsgericht M=A8=B9nchen: HRB 161028 Gescha"ftsf=A8=B9hrer: Thomas Schro"der, Wolfgang Engels Vorsitzender des Aufsichtsrates: Martin Ha"ring *************************************************************************= ****