While I appreciate the limitations of funds and of technology choices, these -
by definition - limit your synchronization choices and will probably impose a
set of reasonably hard and quantifiable limits on how close you can get to
"realtime". This can be modeled but only with intimate knowledge of the
application, the environment, and its inner workings.
There have been some good suggestions for investigation that have been
mentioned in prior posts.
"almost always changing" and "a few seconds out" and "critical" are
unfortunately not quantified terms. Having a handful of dynamic files get
written to your primary disk every minute with the ability to take a minute to
sychnronize them to the secondary with only the implication that people may
have to wait 60 seconds to get their files is decidedly different than having
thousands of files moving around (immediate disk I/O implication, nevermind
application locking and scaling issues for both the main app and the
synchronization process) with legal implications.
Get this sorted out and back it with data points and you'll likely solve your
own problem - which is whether you need a Symmetrix with SRDF down through to
a set of rsync processes to a rewrite of your app to write the files
consecutively to primary then the secondary.
Previous message: [thread] [date] [author] Next message: [thread] [date] [author]
Messages in current thread:
Re: Need some information..., Jason George, (Wed Aug 27, 6:57 am)