Looks more like 51,282% ??
Nick DiLauro
Inovis Inc.
-----Original Message-----
From: MQSeries List [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED]
Sent: Monday, April 17, 2006 12:23 PM
To: [email protected]
Subject: Re: sizing disk space on AIX
That is the MQMD overhead when the message size is only 1 byte...
[EMAIL PROTECTED]
A.CANON.COM To:
[email protected]
Sent by: MQSeries List cc:
<[EMAIL PROTECTED] Subject: Re: sizing disk space
on AIX
WIEN.AC.AT>
04/17/2006 02:45 PM
Please respond to
MQSeries List
Philip,
Thanks for the table. But, I'm a little bewildered by the MQMD overhead
percentage, e.g. 36400%?
[EMAIL PROTECTED]
E.COM
Sent by: MQSeries List To
<[EMAIL PROTECTED] MQS
IWIEN.AC.AT> ERI
ES@
04/17/2006 02:39 PM LIS
TSE
Please respond to RV.
MQSeries List MED
<[EMAIL PROTECTED] UNI
EDUNIWIEN.AC.AT> WIE
N.A
C.A
T
cc
Subject
Re:
siz
ing
dis
k
spa
ce
on
AIX
This may be helpful...
This was tested for Sun, but I'd doubt it would be very different on AIX.
If the messages sent were not removed, and the linear logs were not
compressed, then this chart would give you a good idea of the amount of
disk space you would need. Of course, reclaiming the space depends upon
the behavior of the applications and the frequency your log clean up
utility is run. A couple of other points... we give each queue manager its
own file system. We alert whenever the file system exceeds 80%. Also, if
the queue is empty, and you issue a clear queue, then WMQ will release all
but 1620 bytes back to the OS for reuse. Queue space is allocated and
grows as long as the queue depth and message sizes grow. The queue space
is not normally returned to the OS until the queue is cleared or a qmgr
recycle occurs. It may also release queue space after a WMQ overall
checkpoint occurs. (not too sure when this occurs...).
(Embedded image moved to file: pic05829.jpg)
Robert
Broderick
<[EMAIL PROTECTED] To:
[email protected]
COM>
cc:
Sent by: MQSeries List Subject: Re: sizing
disk space on AIX
<[EMAIL PROTECTED]
WIEN.AC.AT>
04/17/2006 01:58
PM
Please respond
to
MQSeries
List
That would be goo if you had two applications and both sent in the same
amount of messages per hour/day/week/month. What you need to do is figure
out what each application will send in numbers and size. Per QMGR. THis
would include both intr and extr QMGR messages. Now if your QMGR and
application are humming along perfectly you could basically say How much
space do I need within one hour. Ahhhh but this isn't a perfect world. So
now you have to start considering what happens when some bonehead for the
receiving application puts in a code change and the pitcher throw the ball
and there is no one there to catch it. How long will it be OK for this
condition to exist. Will the pitcher keep throwing the ball. You could end
up drowning the umpire!! Now, what happens in a failover situation. Things
have to move and it will be a good bet ALL the applications will not come
up perfect. You need somewhere to store these messages. What do the SLA's
indicate. I will tell you. In a DR there will be a lot of yelling later on
but you can use your SLA in the bathroom as paper.
So what I usually do, and everyone is different, but the main idea is to be
safe. I figure out what a day would cost me in bytes. I multiple that by
two for matrixes' that were not correct and then by 2 for two days worth of
work. If this comes in nice you will have plenty of room to grow.
Oh....that was something else I forgot to mention!!!! Once you buy the
storage it is static in cost. But you do not want to go back to
infrastructure 6 months (if you can last that long) and tell them you need
to purchase MORE hardware. Hit them hard and hit them fast I always say!!!
Just my 1.588798566354373389 cents
bobbee
Date: Mon, 17 Apr 2006 12:35:19 -0400
From: [EMAIL PROTECTED]
Subject: sizing disk space on AIX
To: [email protected]
It's more of a general question regarding sizing disk space on unix.
Unlike the mainframe, where a couple of formulas are provided to help you
size the disk space needed for a queue manager, there isn't anything like
that for unix. What are others doing? Are they simply taking the average
size of all messages, slapping on the MQMD and XMIT header, and then
converting that into total bytes required? Instructions for managing your
mailing list subscription are provided in the Listserv General Users Guide
available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html
#### pic05829.jpg has been removed from this note on April 17 2006 by
Philip J DiStefano
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html
