...
Have you checked SMF historical data? Have you noticed trends or patterns?
...
That's the issue.
The service provider has access to the SMF data; I don't.
I have to take the word of somebody
who has a vested interest in getting an upgrade first.
We may need it; I won't make a recommendation wit
On Wed, 15 Jun 2005 00:00:00 GMT, Ted MacNEIL
<[EMAIL PROTECTED]> wrote:
>
>Of course, we're going to be charged.
>It wouldn't be an issue otherwise.
>
>As I said in my original post(s), I want to see the evidence.
>Otherwise, don't talk to me about it.
>
>
I assume we are talking about ESCON. Ac
en used at
several large sites to successfully help with I/O elimination.
-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Knutson, Sam
Sent: Wednesday, June 15, 2005 SYSN 6:06 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Disk I/O tuning still possibl
ally a good thing.
Regards,
Gary Diehl
-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] Behalf
Of Ted MacNEIL
Sent: Sunday, June 12, 2005 7:00 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Disk I/O tuning still possible?
...
So, I ask, is I/O
...
Is he trying to charge you for them? Ask him how he knows that you
need them.
...
Of course, we're going to be charged.
It wouldn't be an issue otherwise.
As I said in my original post(s), I want to see the evidence.
Otherwise, don't talk to me about it.
-teD
(The secret to success is since
In <[EMAIL PROTECTED]>, on
06/13/2005
at 12:00 AM, Ted MacNEIL <[EMAIL PROTECTED]> said:
>Our service provider is attempting to get us more channels into our
>single DASD controller.
Is he trying to charge you for them? Ask him how he knows that you
need them.
>The bottom line is we have no c
...
This is a simple application of Erlang M/M/C model. With 4 servers at 50%
busy, the probability is 0.1739130435 that all servers will be busy.
...
I realise I was (over-)simplifying.
But, the point is, do I need to upgrade my channels when I have NO channel busy
delay.
And, NO evidence.
-t
>As far as I remember, you have to start digging into 'Eigen numbers' or
>something like that
Yes. It is called the Erlang-C formula. The phone industry started to use
it 80 years ago. Erlang was the pioneer of traffic theory. Just Google
around.
Regards,
Robert Hamilton
: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Disk I/O tuning still possible?
"Robert Hamilton" <[EMAIL PROTECTED]> wrote in message
news:<[EMAIL PROTECTED]>...
> >> So, I ask, is I/O tuning still possible, or even necessary?
> >> ..
> >> I believe you can run them
As far as I remember, you have to start digging into 'Eigen numbers' or
something like that
-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Vernooy, C.P. - SPLXM
Sent: 14 June 2005 08:05
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Disk
"Robert Hamilton" <[EMAIL PROTECTED]> wrote in message
news:<[EMAIL PROTECTED]>...
> >> So, I ask, is I/O tuning still possible, or even necessary?
> >> ..
> >> I believe you can run them quite high, well over 50%.
> >> Since (the simple equation), the odds of all four being busy is:
> >>
> >> 5 *
>> So, I ask, is I/O tuning still possible, or even necessary?
>> ..
>> I believe you can run them quite high, well over 50%.
>> Since (the simple equation), the odds of all four being busy is:
>>
>> 5 ** 4 = .125
>>
>> Or 12.5% of the time there would be no channel busy.
>> Of course, you have to
"Ted MacNEIL" <[EMAIL PROTECTED]> wrote in message
news:<[EMAIL PROTECTED]>...
> ..
> So, I ask, is I/O tuning still possible, or even necessary?
> ..
>
> I've got a related question.
> Our service provider is attempting to get us more channels into
> our single DASD controller.
> We have four.
>
...
The rule of thumb I've heard is that you shouldn't run "channel busy + director
port busy" over 50%.
...
I've run up to 65-70% with no problem.
Which is why I'm asking for evidence.
...
I'd say if you can go to 8 channels in width, without incurring an unacceptable
cost, go for it.
...
time so extra width is generally a good thing.
Regards,
Gary Diehl
-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED]
Behalf Of Ted MacNEIL
Sent: Sunday, June 12, 2005 7:00 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Disk I/O tuning still possible?
...
So, I ask
...
So, I ask, is I/O tuning still possible, or even necessary?
...
I've got a related question.
Our service provider is attempting to get us more channels into
our single DASD controller.
We have four.
We have no Channel Busy Delay.
We have more than adequate response time.
They have supplied no
In a message dated 6/13/2005 4:47:56 A.M. Central Daylight Time,
[EMAIL PROTECTED] writes:
> Then along came RAID arrays.
>
You skip one historic event: Cache in control units
Other major events that have greatly muddied the waters:
(1) the evolution of the Channel Measurement Data. Wi
On Mon, 2005-06-13 at 10:52 +0200, Marian Gasparovic wrote:
> I don't know about other vendors, but on IBM disk arrays
> you cannot have CKD (mainframe) and FB (unix, ;inux, windows etc)
> disks on the same physical disks.
This is supported on STK Icebergs (um, "Sun SVAs").
--
David Andrews
A. D
-- snip --
Then along came RAID arrays.
.
.
.
So, I ask, is I/O tuning still possible, or even necessary?
-- snip --
Yes, even though most people don't bother. When your I/O response time and
throughput are satisfactory, one tends to just leave it as it is.
You can pull RMF reports and look at th
"John Compton" <[EMAIL PROTECTED]> wrote in message
news:<[EMAIL PROTECTED]>...
> (I don't know if this has been discussed before - tried to search the
> archives, but either there's nothing there, or I chose bad search
> arguments).
>
> Back in the days of old (when sysprogs were bold, etc.) we'd
>
> Yet further: Becuase of the enormous amount of storage available in these
> RAID arrays, many companies hold _all_ their disk space requirements
> within an array. Thus, you get disk space for the mainframe, the mid-range
> systems and sometimes even the PCs all mixed up together on a physical
(I don't know if this has been discussed before - tried to search the
archives, but either there's nothing there, or I chose bad search
arguments).
Back in the days of old (when sysprogs were bold, etc.) wed spend many
happy(?) hours tinkering with the physical placement of files in order to
tune
22 matches
Mail list logo
