#198: Consider moving to databases from files for calendar data stores
---------------------------------+------------------------------------------
Reporter: [EMAIL PROTECTED] | Owner: [EMAIL PROTECTED]
Type: Enhancement | Status: new
Priority: 2: Expected | Milestone: CalendarServer-2.x
Component: Calendar Server | Severity: Other
Resolution: | Keywords:
---------------------------------+------------------------------------------
Comment(by [EMAIL PROTECTED]):
Files are a challenge in a couple of ways:
* open/close of lots of small files adds up, and can be very expensive
for reporting.
* we mitigate this somewhat by using SQLite indexes
* Atomicity is pretty well impossible: if I change a meeting with 100
attendees, in the current code, I'm touching 100 files. If an error
occurs mid-way, we have inconsistent data which needs cleanup.
I'd also like to see that for a given meeting (UID = x), that there is one
master copy of the meeting, and attendees have that master copy plus their
own edits. That's possible with files as well as with a DB, but my gut
feel is that it will be harder with files.
--
Ticket URL: <http://trac.calendarserver.org/ticket/198#comment:1>
CalendarServer </>
HTTP/WebDAV/CalDAV Server
_______________________________________________
calendarserver-dev mailing list
[email protected]
http://lists.macosforge.org/mailman/listinfo.cgi/calendarserver-dev
