On Sun, 16 Jun 2013 02:12:56 -0700, Mehrshad Manshadi wrote:

>Hello all,

�does anyone have a "rule of thumb" as to how much Replication for disaster 
recovery slows down overall disk I/O?

"Don't overload the link(s) - or the cache".
First time I saw problems was when the target controller "went away".  The 
source controller kept accepting I/O for (quite) a while then started issuing 
warnings that it was slowing down I/O's. Eventually stopped accepting them at 
all IIRC. All very civilised.

Recently, I observed a situation where a customer started mirroring a new 
second tier module - and all hell broke lose. No messages, no warning, but I/Os 
(mainframe and SAN) went into seconds. Yes, plural.
That's the sort of situation where you need to worry about the process 
priorities Radoslaw mentioned - the defaults seem to be inane as far as real 
world sites are concerned IMHO. But for everyday processing, shouldn't be an 
issue.

Shane ...

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to