Link to home
Start Free TrialLog in
Avatar of peter
peterFlag for United States of America

asked on

websphere

Hello,
I have a requirement to build two servers in Amazon AWS hosted environment to host IBM's Websphere amd MQ instances.
1. Can these be built on Amazon OS 64 bit instances, or do they have to be built on RHEL ?

2. Can anyone recommend approximate file system sizes or point me to a how-to article to architect the sizes of logical volumes to allocate to these instances? I am just not sure how large to make the root file system and which directories to build as logical volumes, so the server does not run out of space down the road.

Here are recommendations I have found:
For Linux, the standard WMQ product binaries placement is in /opt/ibm/mqm while in AIX the product binaries are placed in /usr/mqm. The MQ instance and configurations is placed in /var/mqm.

As shown in the CPA's AIX filesystem partition above, the /var/mqm, /var/mqm/log, /var/mqm/errors and /var/mqm/qmgrs are placed in separated partitions. This is standard practices to allocate separate filesystems for logs and errors.

The partition sizes above are for your reference as the AIX WMB server serves a lot of applications, so the log and error directories are allocated a larger size for log keeping. You may not need 10GB for log space, but do provide ample of space for trouble shooting and it can fill up quickly.

Thank you for any insight in advance!
Avatar of Gary Patterson, CISSP
Gary Patterson, CISSP
Flag of United States of America image

1. Can these be built on Amazon OS 64 bit instances,

Not sure what that means.  Websphere comes in different editions, each with different requirements.  This document describes in detail the system requirements 9including OS support) for each:

http://www-01.ibm.com/support/docview.wss?uid=swg27006921

As for number 2, disk sizing for Websphere and WMQ is completely application-dependent.  I'd need to know a lot more about the applications you will be running and processing volume to even make a guess.  Same applies to log volumes, etc.  How much log volume do you generate, and what are your retention requirements.

Typically, you'd set up a development or test server, and determine proper disk sizing by running your anticipated volumes (or a fraction of anticipated volume and extrapolate).
Avatar of peter

ASKER

Thanks for your reply Gary,
I have no more information unfortunatly on this, what it really funny is the IBM guy that I also was not forthcoming on the file space requirements.
So I have no idea how this server will be used or what i/o it will generate.
It will be a DEV server so we can extrapolate from it's performance.
I am just looking for a ballpark architecture schema to get this started.
So would ballpark be, to build a 50 gb root volume and 3,  20 gb lvm's for /var/mqm/log, /var/mqm/errors and /var/mqm/qmgrs ?
Thanks again,
ASKER CERTIFIED SOLUTION
Avatar of Gary Patterson, CISSP
Gary Patterson, CISSP
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of peter

ASKER

yup, totally agree, and thank you for your help!
I made is 80 gb and 20 gb lvm for logs, will adjust from there.
Avatar of peter

ASKER

Excellent and timely feedback, very much appreciated.