Hi, Clustering adds failover and reliability, but usually doesn't increase performance. Accordingly, if I were you the first setup I'd try is server A for CMS/SQL, and server B for one instance of Tomcat (unclustered). Run a stress test tool at that and see how it responds. If the response is satisfactory, you're in a good spot.
At that point you can add clustering and re-test: if the performance is still good, now you have good performance *and* failover. If not, you might have to go with one Tomcat (unclustered) for performance sake. The important thing, and I know you know it but I'm repeating it for the sake of others, is to benchmark, benchmark, benchmark. Theoretical discussions are only that, theoretical ;) Yoav Shapira http://www.yoavshapira.com >-----Original Message----- >From: Allistair Crossley [mailto:[EMAIL PROTECTED] >Sent: Friday, December 03, 2004 11:10 AM >To: Tomcat Users List >Subject: 2 Servers, 1 CMS, 1 Database Server + ? Tomcats ... > >Hi > >We are running 1 instance of Tomcat on a 2 yr old dual processor P3 1Ghz >with 1GB RAM. The application is an intranet, and it integrates with SQL >Server (running on the same server) via Tomcat-managed JNDI pools. The >application has a business logic layer that has to cope with a mixture of >content management system interactions (also running on same server - >Stellent), direct JDBC calls, and lots of applications use Hibernate ORM. >This server has to take a battering from our 450 staff globally. > >A weeny server for such an application you will probably agree from some of >the specs I have seen on this list. One thing it has done of course is >force me to learn a lot about caching, some profiling and stored procedure >optimization as well as Java code optimization. I remember when programmers >had to alsways take scarce resource into consideration but nowadays things >are just BIG, and I wonder how much of that is attributable to bad coding. > >Anyhow, that aside, my optimizations have only gotten us so far. For the >most part things nip along quite swiftly, but we also see 15-20s response >times say, when an announcement goes out with a link in an email to a >content page in the application. > >So, I finally managed to convince management to buy us a new server. I >believe it will be dual processor Xeon with hyperthreading etc.. but it >should be quicker than our current server. However, I want to utilize the >current server still of course. > >So in my mix I have Server A with competing apps > >1 x CMS >1 x SQL Server >1 x Tomcat > >I am going to start thinking about the best setup for using both servers. >In particular I am interested to hear whether my current idea is any good >or holds, that is > >Keep Server A running CMS/SQL Server. Then use Server B and setup a bunch >of Tomcat instances with clustering. My theory which could be wrong, is >that a bunch of Tomcat's on 1 server is better than 1 Tomcat? Now, I know >this leaves the possibility of some server dying and so on, so maybe leave >the 1 Tomcat on Server A and have a bunch on Server B. > >I am just trying to figure out with these 2 servers, one which will be much >faster than the other, what kind of setup with Tomcat will yield the most >performance for response times to users. > >Any advice appreciated. > >Allistair > > ><FONT SIZE=1 FACE="VERDANA,ARIAL" COLOR=BLUE> >------------------------------------------------------- >QAS Ltd. >Developers of QuickAddress Software ><a href="http://www.qas.com";>www.qas.com</a> >Registered in England: No 2582055 >Registered in Australia: No 082 851 474 >------------------------------------------------------- ></FONT> > > >--------------------------------------------------------------------- >To unsubscribe, e-mail: [EMAIL PROTECTED] >For additional commands, e-mail: [EMAIL PROTECTED] This e-mail, including any attachments, is a confidential business communication, and may contain information that is confidential, proprietary and/or privileged. This e-mail is intended only for the individual(s) to whom it is addressed, and may not be saved, copied, printed, disclosed or used by anyone else. If you are not the(an) intended recipient, please immediately delete this e-mail from your computer system and notify the sender. Thank you. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
