On Tue, Sep 18, 2007 at 03:23:13PM -0700, Jeroen van Aart wrote:
(trying to be a bit oblique so as not to anger/embarrass anyone).
I'd say a healthy dose of embarrassment might help prevent such mistakes
in the future. :-)
I'm sure that they were plenty embarrassed already; the campus user
Dan Pritts wrote:
On Sun, Sep 16, 2007 at 01:33:50PM -0700, Mark Crispin wrote:
My understanding is that this is what caused the spectacular failure of
the new mail system at a big 10 university last fall.
they had bought their fancy NFS servers from a large vendor known for
their PCs.
You
On Tue, 18 Sep 2007, Jeroen van Aart wrote:
don't always pull themselves out from the middle. UW's DNS-based approach
makes that unnecessary.
Is there a place describing that approach? It sounds quite interesting.
Here's an old document that the developer of that system wrote to describe
On Sun, 16 Sep 2007, Bob Atkins wrote:
While I have been acutely aware of the issues you describe below regarding
the shortcomings of NFS and simultaneous imap access I don't see how to scale
a large imap server environment without NFS.
This is something that has been discussed for at least
On Sun, 16 Sep 2007, Mark Crispin wrote:
UW and CMU adopted very different models on how to do scaling to very
large communities. UW uses a DNS-based solution; Cyrus has a facility
they call Murder. Either one is FAR preferable to NFS. UW's solution
is more general as it can be applied to
On Sun, 16 Sep 2007, John Kelly wrote:
Can you speculate on how mix compares to mbx, in terms of I/O load?
Seems like I remember you saying mix does not need atime.
Correct.
In terms of data volume, mix's I/O load is comparable to mbx.
mix wins big in reduced seek times when reading message
On Sun, 16 Sep 2007, David Carter wrote:
While I agree just about everything that Mark has to say about NFS (been
there, done that, got the T-shirt as well), it is probably worth pointing out
that the Cyrus Murder is just a (very clever) IMAP proxy.
Other, generic, IMAP proxies can be used
We run uwash imapd against a traditional unix /var/mail. There is a
mixture of imap clients (thunderbird, pine, Mac mail, squirrel, etc). We
are plagued with people leaving one client going (maybe at home) and
starting a different one at another location (say their office).
This ends up with