Christoph Berendes wrote:
Why so many, in separate files, you ask...
1. why not? (what's the cost/risk of multiple Data.fs files and mount
Because it's not a common use case, so you end up pushing limits and
bumping into bugs that no-one else cares about ;-)
2. makes it easy to move a single client site between instances and
servers (we don't yet know how load is going to mount up, and whether
we'll need to set up a new server at, say, 30 sites, or 100, or ...)
3. site import/export doesn't work as well, and it seems that
import/export of "large" Zope objects is deprecated...
huh? rubbish. What specific problems have you had?
4. I'm told that, occasionally, Data.fs become corrupted.
I've yet to see that in 5 years of extensive ZODB use, other than by
hardware failure, in which case you're screwed however you have things
set up ;-)
We have some
protection if that happens, since it should (I'd think) affect only one
client, not many.
Nah, the stuff you're worried about will trash multiple mounted storages
just as well as one big one ;-)
I'm very interested in answers to #1 (things that will make me regret
our current strategy)
The biggest risk is that you'll be the only person who cares if things
go wrong because of that setup...
Simplistix - Content Management, Zope & Python Consulting
Zope maillist - Zope@zope.org
** No cross posts or HTML encoding! **
(Related lists -