Re: Speed of Block

My thanks to Bogdan and Neil for your replies and for confirming the that the steps that I listed were the required steps.? I discussed the situation again with staff and did a timing test.? The longest part of the process was the actual BCV split which takesplace after the onmode -c block takes ?about 12 to 15 seconds.?It turns out that a portion of the 12 to 15 seconds is due to use of NSF and commands gong across servers.? So, this item is now being put to rest. Thanks again eveyone. Regards, Bruce -----Original Message----- From: Bogdan BOTEZ <bmbogdan@gmail.com> To: brucemarchu@aol.com Sent: Fri, 27 Feb 2009 11:26 pm Subject: Re: Speed of Block On Fri, Feb 27, 2009 at 21:36, <brucemarchu@aol.com> wrote: > Hi Bogan, > > Thanks for your reply.? I will test tonight and will then reply.? I can not > interrupt the processing happening now. > > Also, does anyone know of an alternate method of accomplishing the > equivalent of the procedures listed below? > > Thanks, > Bruce > The rationale behind the question is simple, as theBP also asked: is it the total duration that is bothering you, or it is just the split duration? It is true that maybe each operation is taking some (few) seconds, but the blocking part is (ckpt), (-c block, split, -c unblock). If the checkpoint before is longer, then maybe you can change the configuration to have smaller checkpoint duration (LRUs, mindirty, maxdirty, bufferpool size, etc). As for the steps, these are the required steps - and believe me when I say - this is the fastest recoverable copy. There are people doing splits without blocking, but: * there is no guarantee that the copy instance can recover * is an unsupported operation (with IBM support), don't complain to IBM if it didn't worked. It is important to understand FIRST where is the problem, if there is one. Regards, Bogdan BOTEZ.