Manshadi,

1- How can I find consistency group time and the delay time?
Look at the XQUERY command. I think it would be something like "XQUERY
VOLUME(ALL) STATUS" but check the manual. This is issued against the STM and
it is the consistency time of records applied to the secondary volumes, and
not the latest time in the XRC journal.



2- Why the Batch jobs on primary ESS take double execution time?
I'm guessing you have flow control turned on. A 150Mbs OC3 link is barely
15MB/sec, and eight channels of FICON 8S can write at 5600MB/sec. That makes
your outflow rate around 0.3% of the inflow. This is the worst case and it
would take a few seconds for XRC to reach maximum cache usage and drop the
link. This is called a sidefile puncture. Flow control prevents this by
slowing down writes as the XRC sidefile uses more cache, which in turn slows
down your batch.

You mentioned an ESS, which means you may still be on ESCON. Even so,
8xESCON can write at 136MB/sec, which puts your link capacity at 11% of the
potential write capacity. In simple terms if I write 136MB in one second,
and send 15MB in the same second I have accumulated 121MB of side file. In 9
seconds it is a GB, and in 15 minutes it is 100GB of cache used by the
sidefile. How much cache do you have, and if you are using FICON this
happens much faster.

You also need to provide adequate MIPS in the remote site for the SDM to
read and apply record sets from the Primary storage. This can also cause the
sidefile to grow and accumulate in the primary storage and trigger flow
control.



3- How can I control the delay time?
You need to match the bandwidth between your primary site with your remote
site. Understanding your write rate profile by time of day with XRC
suspended will give you an idea of the peak demand. After that it is good
old capacity planning and perhaps investment in something faster than OC3.

If your ESS is actually a DS8K you may be able to get over the hump by
adding cache, but it still comes back to understanding your unfettered write
rate.


Ron







> -----Original Message-----
> From: IBM Mainframe Discussion List [mailto:[email protected]]
> On Behalf Of Mehrshad Manshadi
> Sent: Tuesday, February 26, 2013 12:25 AM
> To: [email protected]
> Subject: Re: [IBM-MAIN] XRC Consistency Group Time Report.
> 
> Hi,
> Thanks.
> 
> Whenever we are using XRC the mirror is Asynchronous. am I right?
> Is there possible establish synchronous pair with XRC? any parameter?
> 
> Regards
> Manshadi
> 
> 
> 
> ________________________________
>  From: Mehrshad Manshadi <[email protected]>
> To: [email protected]
> Sent: Tuesday, February 26, 2013 11:52 AM
> Subject: Re: XRC Consistency Group Time Report.
> 
> Hi,
> Thanks.
> 
> Whenever we are using XRC the mirror is Asynchronous. am I right?
> Is there possible establish pair with XRC? any parameter?
> 
> Regards
> Manshadi
> 
> 
> 
> ________________________________
> From: Mike Schwab <[email protected]>
> To: [email protected]
> Sent: Tuesday, February 26, 2013 11:44 AM
> Subject: Re: XRC Consistency Group Time Report.
> 
> 2. If you have synchronous mirroring, your job on the primary is waiting
for
> your data to transmit to the secondary and the write to complete and the
> confirmation to transmit back.
> 
> 3. Async does not wait for this to happen.
> 
> On Tue, Feb 26, 2013 at 2:02 AM, Mehrshad Manshadi
> <[email protected]> wrote:
> > Hi,
> >
> > We have an XRC session with 590 volume 3390 model 9, which is
> established on disaster recovery site on 450 k far from our local site.
The
> connection is STM1 150 mbs between local and disaster recovery site.
> >
> > we are using Cisco MDS 9222i as a switch between ESS on local site and
> SDM on disaster recovery site with one channel.
> >
> > I have three question :
> >
> > 1- How can I find consistency group time and the delay time?
> > 2- Why the Batch jobs on primary ESS take double execution time?
> > 3- How can I control the delay time?
> >
> > Best regards
> > Manshadi
> >
> > ----------------------------------------------------------------------
> > For IBM-MAIN subscribe / signoff / archive access instructions, send
> > email to [email protected] with the message: INFO IBM-MAIN
> 
> 
> 
> --
> Mike A Schwab, Springfield IL USA
> Where do Forest Rangers go to get away from it all?
> 
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions, send email
to
> [email protected] with the message: INFO IBM-MAIN
> 
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions, send email
to
> [email protected] with the message: INFO IBM-MAIN
> 
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions, send email
to
> [email protected] with the message: INFO IBM-MAIN

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

Reply via email to