-- snip -- But I was under the impression a multi-site GDPS/PPRC environment is suppose to offer continous availability, how does it do it if both CF are at one site and the site destroyed. If you're using dedicated standalone CF (not ICF), why whould it matter if the CF were local or physically remote ? For all intents and purposes won't it be the same as if it were local except for the performance impact ? Won't the CF duplexing protocol ensure both CF have exactly the same consistent content all the time (2-phase commit) ? The whole idea of duplexing is to allow failover to simplex mode of the remainding CF no ? -- snip --
You must have a CF at each site. In a fail-over situation, the DASD will be consistent. The reason that CF duplexing is not (yet) allowed, is that the structure contents may be ahead of the DASD when the freeze takes place. The DASD I/O is set to long busy in a freeze situation, but that is not the case with the CF I/Os. The current correct manner is to rebuild the structures. I have seen a reference to DB2 and multi-site data sharing. Someone with DB2 knowledge will have to take it from here. John ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
