I believe that clusters, whether active-active or active-passive, provide poor business value. I truly believe that clusters actually reduce availability for most users; the posts in this form help confirm that. The reported list of problems are long but the list of those who believe a cluster has "saved their bacon" (an Ed Woodrick term) is tiny. Those who are considered "experts" on this list, its companion lists, and newsgroups (and take that moniker for whaterver it's worth) nearly unanimously (if not fully unanimously--I can't recall an MVP who thinks they're worth the effort, but there may be one) agree with me.
With two-way clusters, why bother with an active-active at all? You have to scale each machine so that it can handle all of the users, so what do you get by running each node at half capacity? You're just buying two machines where one will do. You should buy the most internally redundant system you can and consider it to be a single-node cluster. Properly configured, Exchange really does fail very rarely, and clusters provide no protection from most failure points. For example, clusters do nothing to reduce any problems related to corruption or disk space management since they share the same database and log files. What does a cluster protect you against versus an internally redundant single-node system? Not much more than a failure of a motherboard. And how often does that happen? Does it justify more than doubling the system cost? Just about everything else can be made redundant. One of the pimary arguments I hear for clustering is that it allows an organization to make its service level agreements. Frankly, if an SLA is written such that the business cost of a planned outage on, say, one Sunday morning a month makes it cost-effective to deploy a cluster, then the business value of the SLA needs to be revisited. Ed Crowley MCSE+Internet MVP Freelance E-Mail Philosopher Protecting the world from PSTs and Bricked Backups!T -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Jason Clishe Sent: Monday, November 24, 2003 2:11 PM To: Exchange Discussions Subject: Looking for anti-clustering arguments OK, I know the majority of this list is against clustering Exchange. But does anyone have some sound reasoning behind this argument that could be used to convince someone at the executive level? I don't think many executives are too compassionate for the "it's harder to administer" argument. And I need more ammo than just saying that clusters won't protect you from database corruption anyway. Thanks Jason _________________________________________________________________ List posting FAQ: http://www.swinc.com/resource/exch_faq.htm Web Interface: http://intm-dl.sparklist.com/cgi-bin/lyris.pl?enter=exchange&text_mode=&lang =english To unsubscribe: mailto:[EMAIL PROTECTED] Exchange List admin: [EMAIL PROTECTED] _________________________________________________________________ List posting FAQ: http://www.swinc.com/resource/exch_faq.htm Web Interface: http://intm-dl.sparklist.com/cgi-bin/lyris.pl?enter=exchange&text_mode=&lang=english To unsubscribe: mailto:[EMAIL PROTECTED] Exchange List admin: [EMAIL PROTECTED]
