Michael Sims wrote:
Quick question on this. If I setup an active/passive cluster and put the mail spool AND all of the application data on a SAN that both nodes have access to (not simultaneously, of course), doesn't that bypass the need for using "mupdate_config: replicated"? Thanks...
This is the setup we run, works fine. At the time we installed we were leery of the Murder architecture anyhow with mupdate server as a single point of failure. There are some denigrate active/passive cluster as "wasteful" of hardware, but frankly hardware is cheap and usually the people who bring it up are BEAN-COUNTERS who will claim ignorance of decisions when the critical service is hosed. I find it difficult to conceive a modern system can be overwhelmed such that there's any meaningful benefit to active/active. I like the active/passive configuration because we can patch the idle node, then switch to it, and if patching broke something we switch back no big deal. Our mail store are ZFS on double-path SAN switches and we've had zero unexpected hardware downtimes despite various hardware failures.
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature
---- Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
