Re: Bugs with mixed Zimbra 7/Zimbra 8 cluster

[Rich Graves <rgraves@carleton.edu>]

Pablo: 
> Let me know when you get to that point of migration . I will send you a perl script that can do the migrations for you. 

Right now would be great, to see what I might be getting into. 

Vlad: 
> We were running 7.1.something. on the old machines. Out entire install (/opt/zimbra) is
> sitting on our Netapps, so this helped as you will see, but you can probably fake it
> with copies. On the new hardware, running RH6, we installed zimbra with -s option (just
> a software install), and then removed /opt/zimbra as just installed 

I did this in July 2011 when we upgraded from RHEL4 to RHEL5. Good to know that you were still able to go from RHEL5 to RHEL6/CentOS6. That's my Plan B.

OK, so here's what's really going on: 

Our server hardware needs replacing. Our SAN needs a substantial upgrade. Since 2007 I have been running one monolithic Zimbra server for simplicity. I know that I need to separate out the MTA and an LDAP replica. But I am getting zero help from other local staff, I need to be doing more strategic things, and we have had zero qualified applicants for https://jobs.carleton.edu/postings/1773 in six months. We have two good options for outsourcing Zimbra, both of which use Ubuntu not Zimbra, at least one of which is an active and valued member of this mailing list. We would get a level of DR not achievable locally, reduce costs, and allow me to focus on more interesting things. We're at a good point in the server and SAN life cycle to jump. I think it makes tremendous sense to run Zimbra at point B. But it is proving very difficult to get it there.
-- 
Rich Graves <rgraves@carleton.edu>
Carleton.edu Sr UNIX and Security Admin