The "IBM Lotus Domino Performance in a VMware Infrastructure 3 Environment" whitepaper was finally approved by VMware and IBM and is now available to customers under Non Disclosure Agreement. This means we aren't allowed to post it on a public website, however if your company has a current NDA agreement with either IBM or VMware, you can contact your representative and request a copy.
Ivan Dell’Era | 16 April 2008 10:39:10 AM ET | | Comments (2) | Permanent Link
We reached an agreement with VMware and they agreed to publish our findings in a whitepaper, however it won't be publicly available as it will be released under NDA (Non Disclosure Agreement). It's in the (hopefully) final rewrite by the tech editor and if no major changes are requested, we should see it in a month (I know, I promised it since last year, but it's coming, really).
Ivan Dell’Era | 8 April 2008 11:06:48 AM ET | | Comments (0) | Permanent Link
The whitepaper for Sametime is available on the VMware web site: http://www.vmware.com/files/pdf/ibm_lotus_sametime_ref_arch_vi3_wp.pdf
Ivan Dell’Era | 8 April 2008 09:43:00 AM ET | | Comments (0) | Permanent Link
We don't have yet an agreement with VMware on publishing the results. I'm working on other options to publish the results but currently we aren't allowed to publish the whitepaper we completed based on the benchmark we completed last Summer.
Ivan Dell’Era | 11 February 2008 03:22:37 PM ET | | Comments (0) | Permanent Link
Several of my colleagues in the IBM Software Services for Lotus Americas team recently had the Lotus Notes / Domino 8 Upgrader's Guide published.
This guide is sure to become an indispensable resource for Notes users, administrators, and developers interested in learning more about what's new in the latest version of Notes/Domino. The pages are packed with useful information on new Notes 8 features and tools, upgrade approaches, integration considerations (including with Lotus Quickr, Sametime, and Connections), and much more.
If you're looking to add this to your bookshelf, it can be ordered from Packt Publishing and Amazon.com.
About the Author(s)
Tim Speed
Tim Speed is an IBM Certified Systems Architect with IBM Software Services for Lotus. In that capacity, he is responsible for designing, implementing, and supporting various engagements with its clients. Mr. Speed lives in Denton, Texas and has been an IBM/Lotus employee for over 12 years in a variety of networking, technical, hardware and software support and consulting positions. He has been working with Notes for over 15 years focusing on administration roles and infrastructure. He also has international experience with working on infrastructure engagements in Spain, Japan, Hong Kong, Singapore, Malaysia, the UK, and Indonesia.
Brad Schauf
Brad Schauf is an IBM Executive I/T Architect with over 20 years of experience in the computer services and consulting industry. He has experience with enterprise-wide software and messaging and portal deployments, with a concentration on Lotus Notes/Domino messaging infrastructure architecture, application development, and integration as well as WebSphere portal architecture design and deployments. His experience includes API-level application development and lead programmer, enterprise lead for messaging and portal deployments to General Manager including P&L commitments. He was a founder of a successful IBM business partner before joining IBM in 1999.
David Byrd
David Byrd is an IBM Senior Certified Executive IT Architect with IBM Software Services for Lotus from Fayetteville, GA. He has been an IBM/Lotus employee for over 9 years in a number of consulting positions covering various technology areas. David has a deep background in virtually all areas of Lotus products and technologies covering areas ranging from low-level API development and collaborative application architectures, to security architectures and messaging architectures. His current focus is on Lotus Quickr as well as other team collaboration technologies and its deployment within enterprise customers. He has worked with Lotus Notes and Domino for over 15 years.
Joseph Anderson
Joseph Anderson is an IBM Certified Senior Managing Consultant from the IBM Software Services for Lotus team. Joseph has worked with Lotus Notes/Domino, Lotus Sametime, and Lotus QuickPlace since the early 1990s, primarily as a consultant. He is currently working with the Competitive Software team focusing on Domino/Notes administration, migration/upgrade, and security. Prior to working in the consulting industry, Joseph worked in the legal industry as a Director of Operations, where he leveraged his Master of Science in Legal Administration from the University of Denver College of Law.
Barry Rosen
Barry Rosen is currently an Advisory IT Specialist with IBM Software Services for Lotus. During the last two years, Mr. Rosen has worked on several large messaging and migration projects as well as performed Domino upgrades and messaging assessments. Before that he was a Software Engineer in Lotus Support for over five years. While in support Mr. Rosen was on several teams specializing in mail routing, Lotus Notes Client, calendaring and scheduling, and server core. He focused on clustering, Lotus Notes for the Macintosh, and rooms and resources. Currently Mr. Rosen resides in Houston, Tx. Having graduated from the University of Texas at Austin, Mr. Rosen enjoys following Longhorn sports.
Bennie Gibson
Bennie Gibson is an IBM Certified Systems Architect with IBM Software Services for Lotus. In that capacity, he is responsible for managing various engagements with its clients. Mr. Gibson lives in Wake Forest, NC and has been an IBM/Lotus employee for over 24 years in a variety of sales, consulting, and management roles. He has been working with Notes for over 10 years focusing on architecture and infrastructure. He also has international experience with working on infrastructure engagements in Malaysia.
Dick McCarrick (Editor)
Dick McCarrick is a freelance writer who has worked extensively with Lotus Notes and Domino over the years. Dick spent over 15 years with the Lotus Notes and Domino team, initially as a documentation writer, then later with developerWorks: Lotus. Since leaving IBM, he continues to be involved with Notes/Domino, co-authoring three previous books on this product.
Stephen Hardison (IBM Official Reviewer)
Stephen Hardison is an IBM Certified I/T Specialist with IBM Software Services for Lotus. He focuses on the design, implementation, and assessment of large-scale collaborative solutions based on Lotus Domino, WebSphere Portal, and Lotus Connections. Mr. Hardison joined IBM in 1999, and has worked in the Information Technology industry for over 20 years. He has worked on several world-wide deployments of Lotus products. Customer engagements have taken him to Argentina, the Bahamas, Brazil, Canada, France, Switzerland, and the United Kingdom. He lives near Austin, Texas.
David Bell | 23 January 2008 05:43:37 PM ET | Orlando, FL | Comments (1) | Permanent Link
What is a best practices approach to supported Portal versions in a deployment?
I've been asked what is a best practices approach to support WebSphere Portal and WebSphere Application Server versions in a typical production environment.
Here's the simple (but important) answer:
Although other configurations and versions may be supported, my advice is:
Stick to the suggestions made on the WebSphere Portal Server Support Page: http://www-306.ibm.com/software/genservers/portal/support/" title="http://www-306.ibm.com/software/genservers/portal/support/">http://www-306.ibm.com/software/genservers/portal/support/
The section "Recommended fixes and updates" contains quality recommendations about version dependencies between WebSphere Portal, WCM, WebSphere Application Server and WebSphere Process Server.
It is a best practice approach to regularly check the Support Page for important updates.
If you plan to update a certain component of your deployment, check the dependencies first and download other necessary fixes, if applicable.
Markus Nagel | 22 January 2008 05:42:39 PM ET | | Comments (0) | Permanent Link
Integration between Portal and Quickr (J2EE Version)
Question: We've just had a customer asking the different integration approaches between Portal and Quickr (J2EE).
1.) The easiest approach for a new Quickr installation is: Why not use Quickr as your Portal? Quickr (J2EE) lives on a full-fledged WebSphere Portal 6 installation - so you could use this as your portal and deploy your applications onto the Quickr Portal or even use Virtual Portals.
The System Verification Test (SVT) Team has published some scenarios with Quickr8 Services for WebSphere Portal. http://www-1.ibm.com/support/docview.wss?uid=swg27010312
2.) If you want to have separate Portal Servers - you could use WSRP Portlets to surface your portlets on the Quickr side where needed. To keep the setup easy, you might want to use the same URL domain (for the LTPA token) and the same LDAP server. Integrating these is straightforward.
Slow Login Performance when using Active Directory
Question: How do I increase login performance when using Microsoft Active Directory as LDAP server?
Firstly, I'd like to stress the fact that the low login performance is not a problem that is specifically related to WebSphere Portal. Any LDAP based system (your Apache Server, PHP Application or any other tool using LDAP as authentication source) will have the same problem. This is related to the hierarchical structure that can grow to deeply nested subtrees in large deployments. When WebSphere Portal (or any other LDAP-based authentication mechanism) issues a search request, Active Directory has to search for the users through all sub-trees recursively, which can be time-consuming.
Depending on the Active Directory setup, maybe some of the sub-trees reside on different servers which results in network traffice between AD server nodes.
There a some ways around this:
1.) Active Directory provides "Global Catalog" Servers with a condensed version of the directory with better response times (Windows and AD internally use these, too - but with proprietary protocols, not LDAP )
You should consult with your Windows / AD admins how and where these "GC" servers have been set up. Specifically use these, not just any AD server.
2.) If it is not possible to use the GC servers, use Tivoli Directory Integrator to create a condensed version of your directory for LDAP authentication purposes only. If you do, be aware that changes made to user profiles (e.g. through self-care) need to be replicated back to the Active Directory.
3.) Re-Design the Active Directory (which is not possible in most of the cases, but for sake of completeness...) to store users and groups at specific places / in specific trees instead of scattered throughout the AD infrastructure.
Markus Nagel | 22 January 2008 05:35:19 PM ET | | Comments (0) | Permanent Link
Attached you will find the soft copy. For the many people who have asked Joe and I "Will there be a repeat of this session?" please note this to the Lotusphere staff and through Lotusphere Online.
Please click the title to download.
Enjoy
Elliott Harden | 22 January 2008 02:06:06 PM ET | | Comments (1) | Permanent Link
In the NSD Hands On Lab many customers asked how do we get detailed functions of our customer application in the NSD output. When NSD runs it annotates call stacks against LotusDomino.sym (symbol file). This produces functions and offsets to give further insight to what when and how the process crashed. After talking with Jim Rouleau when NSD runs it will look through all *.sym files in the Notes/Domino program directory. Now you can take the map file from your custom application, create a sym file from your map file. Once you have a custom sym file you can now get your custom applications annotated data in the NSD.
For more information check out Map2iSym.exe located in the CAPI Toolkit
http://www-12.lotus.com/ldd/doc/tools/c/7.0/api70ug.nsf/85255d56004d2bfd85255b1800631684/00cb00910049008485255e3d00774c2f?OpenDocument
Elliott Harden | 22 January 2008 11:31:56 AM ET | Orlando, FL | Comments (1) | Permanent Link
Lotus Domino 8 server performance
With Notes / Domino 8 we have an all new, redesigned Notes client with increased functionality, and a far superior user interface. Maintaining Domino mail server scalability with this new client was of critical importance, and we hope to show you in this article that Domino 8 not only equals, but generally outperforms the previous release.
In this article we will show how Domino 8 with the Notes 8 client performs across a variety of server platforms. We will show tests with thousands of simulated Notes users running thru common mail and calendar scenarios. Since the new client has changed some of the requests made of the Domino server, we created a new workload (N8Mail) that models transactions observed while tracing the Notes 8 client as a real Notes 8 user performed the chosen tasks. Then we did the same thing for the Notes 7 client, and created the N7Mail workload which models the same user actions, but with the set of requests that a Notes 7 user would generate.
This article is the full discussion of the Domino 8 Performance results:
http://www.ibm.com/developerworks/lotus/library/domino8-performance/index.html?S_TACT=105AGX13&S_CMPíU
The network bandwidth increase, for D8 when compared with D7, reported in The D8 server performance with the Notes Client report,
was due to 2 reasons.
- A NotesBench N8 Client workload simulation bug. The N8Mail workload in Notesbench 8.0 Gold version had two issues.
Issue #2 was that the workload simulated a full Note Open instead of summary only Note Open call during the delete transaction in the workload.
These two notesbench bugs in the N8Mail workload caused the server to return more data to the notesbench clients than what the Notes 8 client would retrieve for the above two transactions in the N8Mail workload .
- The Notes 8 client retrieves more documents in a view for a given screen size, than the Notes 7 client. This was a design consideration in the Notes 8 Client to cache more documents for a better user experience during page scrolling . The Notes Client uses the NIFReadEntries API call to fetch the entries in a Notes View . The Notes 8 clientretrieves 50 entries in the Inbox , while Notes 7 Client retrieves only 42 entries in Inbox during .
For D8.0.1, the NotesBench N8 Client workload simulation bugs were fixed. The number of read entries the N8 client does was not changed.
Below is a chart that illustrates the effect of the Notesbench workload change to the Network Bandwidth calculation. Note, the remaining increase is a result of the new to N8 readentries increase. Other metrics, such as CPU and disk I/O did not change with the workload change.
Network Bandwidth Resource Usage at 4000 users for AIX
| Domino 7 | Domino 8 | % Change | |
| Mail Template | Mail7.ntf | Mail8.ntf | n/a |
| Workload | N7Mail | N8Mail | n/a |
| Network Kbytes/sec (with old N8Mail workload ) | 2,427 | 2,903 | 20% |
| Network Kbytes/sec (with new N8Mail workload ) | 2,386 | 2,550 | 7% |
Network Bandwidth Resource Usage at 4000 users for Windows
| Domino 7 | Domino 8 | % Change | |
| Mail Template | Mail7.ntf | Mail8.ntf | n/a |
| Workload | N7Mail | N8Mail | n/a |
| Network Kbytes/sec (with old N8Mail workload ) | 2,632 | 3,072 | 17% |
| Network Kbytes/sec (with new N8Mail workload ) | 2,540 | 2,768 | 9% |
Andrew Nolet | 22 January 2008 09:22:08 AM ET | | Comments (0) | Permanent Link
It happened again. I was presenting best practices for upgrading to version 8 to clients and business partners on Monday and then I talked about ODS. Suddenly, I noticed that everyones eyes glazed over, as they weren't really sure what this TLA (Three Letter Acronym) was all about. As I suspect that more than just these esteemed people might not know for sure what ODS is and what it all has to do with Notes and Domino, let me take a moment to expand upon this as part of this blog entry.
On-Disk Structure (ODS)
Every major release of Notes/Domino (e.g., versions 2, 3, 4, 5, 6/7 and now 8) has included significant architectural changes to the database structure, also known as the On-Disk Structure (ODS). These architectural changes enables the Notes and Domino development team to improve performance and scalability in each release. Upgrading the databases provides a lot of benefits, with very low risk, so this has always been an important early step in any Notes and Domino upgrade plan.
The ODS version for each major release of Notes and Domino is as follows:
- v2 -- ODS 16/16.3
- v3 -- ODS 17/17.3
- v4 -- ODS 20
- v5 -- ODS 41
- v6/v7 -- ODS 43
- v8 -- ODS 48
So there is now a new ODS for Notes and Domino 8, which is ODS 48. However, that ODS version is not enabled by default for new or upgraded applications. This permits an easy migration from 6.5 and 7.0 environments.
So what does ODS 48 provide?
A major consideration for mail servers when considering upgrading to a newer version of Domino is the impact this will have on disk space utilization resulting in the use of the new mail template. Typically, new templates imply new features and new functionality and these generally require more space by the newer design contained in the template. With version 8, it is no exception. There is considerably more features offered to the user and these will consume more space. The Notes Mail 7 Template requires 17 GB of disk space (for the design alone) whereas the Notes Mail 8 Template requires 25 MB (both based on ODS 43).
Thus, if the design is applied to 1,000 mail files present on the server, there will be an increase of 8GB in disk space utilization, for a total of 25GB just for the design alone. Given that many servers handle more than 1,000 users (and a commensurate number of mail files), this may have a broader impact on disk space utilization.
So ODS 48 helps by enabling design note compression. This is an option by which the design notes within a database can be compressed as part of the new On Disk Structure in Notes and Domino 8. This feature reduces the size of databases by compressing design elements by up to 60%. For example, the design elements in a database based on the standard Notes 8 Mail template (Mail8.ntf) reduce in size from approximately 25MB to 11MB once design compression is implemented. The compression percentage achieved will vary from database to database based upon the individual compression ratio achieved for each design element in each database.
In real terms, it means that, upgrading to Domino 8 and ODS 48, there would be a decrease in disk space used just by upgrading instead of the customary increaase.The decrease, for 1,000 mail files would actually be of 6 GB, significantly decreasing the disk overhead on a per-mail file basis.
How is ODS 48 implemented?
As mentioned, ODS 48 is not enabled by default for new or upgraded applications. ODS 43 is the norm. To implement ODS 48, the parameter CREATE_R8_DATABASES=1 needs to be added to the notes.ini file to enable ODS 48 (Version 8) as the default.
Thus, for the Domino 8 server, applications residing on the server are upgraded in the following manner:
- The CREATE_R8_DATABASES=1 notes.ini parameter needs to be added to the notes.ini file on the Domino server (note: the parameter is dynamic and can be enabled with 1 or disabled with 0 using "set config"; as well, the parameter is not case sensitive);
- Load Compact -c against the applications that need to be upgraded to ODS 48 (note: if an application is updated by error, it is possible to revert the ODS to the previous release with Compact -r).
Any new applications created on the Domino server will be created with ODS 48, provided that the CREATE_R8_DATABASES parameter is set to 1. If the value is returned to 0 (zero), new applications will be created with ODS 43. The procedure is the same for the Notes client.
Frederic Dahm | 18 January 2008 12:46:25 AM ET | Montreal, QC | Comments (2) | Permanent Link
Maintaining IBM Lotus Domino V8 mail server scalability with the IBM Lotus Notes V8 client is critical, and this article shows that Lotus Domino V8 not only equals, but outperforms the previous release. Let us show you how Lotus Domino V8 with the Lotus Notes V8 client performs across a variety of server platforms. Stop into the DPI lab and ask for Joe, Harry, Andy and Raz. We can help you better understand Domino 8 server scalability.
http://www.ibm.com/developerworks/lotus/library/domino8-performance/index.html?S_TACT=105AGX13&S_CMPíU
Joe Malek | 16 January 2008 11:22:35 AM ET | | Comments (0) | Permanent Link
A thing that comes up frequently when talking with customers is where can they find information about Lotus Software products, other than in the online help databases and the Redbooks/Redpapers? Well, let's take a moment to look into that...
Much like other IBM/Lotus products, Domino has joined the Information Center way of delivering help information. Let's not worry too much, there is still a place for .nsf files. After all, the "Lotus Domino 8 Administrator Help" database ships with the product and is also available via the Lotus Domino Documentation site. But there is indeed a new addition to the Domino Documentation site: the InfoCenter.
The "Lotus Domino 8 Information Center" contains the same information as that of the "Lotus Domino Administrator Help" database, but in a different format. The format matches that of the InfoCenters for Lotus Sametime, Lotus Quckr, Lotus Connections and WebSphere Portal. So, now you have a choice of format when you refer to Domino help.
Provided below is the list of links where you can find the documentation you need on-line, over the World Wide Web, through an Internet connection.
- Lotus Domino documentation - the main Lotus Domino documentation site: http://www.ibm.com/developerworks/lotus/documentation/domino/
- Lotus support: Self-help and services (includes access to the Lotus support knowledge base): http://www.ibm.com/developerworks/lotus/support/
- IBM Lotus Domino and Notes Information Center: http://publib.boulder.ibm.com/infocenter/domhelp/v8r0/index.jsp
- IBM Lotus Sametime Version 7.5.1 CF1 information center: http://publib.boulder.ibm.com/infocenter/sametime/v7r5m1/index.jsp
- IBM Lotus Quickr Version 8.0 Information Center: http://publib.boulder.ibm.com/infocenter/lqkrhelp/v8r0/index.jsp
- IBM Lotus Connections information center : http://publib.boulder.ibm.com/infocenter/ltscnnct/v1r0/index.jsp
- IBM® WebSphere® Portal, Version 6.0 Information Center: http://publib.boulder.ibm.com/infocenter/wpdoc/v6r0/index.jsp
Frederic Dahm | 16 January 2008 12:45:54 AM ET | Montreal, QC | Comments (0) | Permanent Link
The Domino server (and a majority of the Lotus Software products) are easy to install. Generally, all it takes is to launch the installation executable and answer a few questions (e.g., where to install the product, what type of install will it be, etc...). Afterwards, there are some more questions as to how to properly configure the product (e.g., what are some of the basic parameters for it to work properly). Once this is done, the product in question will generally do what it is supposed to do and provide the services it is meant to provide.
But that's only part of the story. If it were the whole story, there would be the "Meet the Developers" lab and that would be it.
Thing is, the products are not islands in of themselves and they need to interact with the rest of the infrastructure, such as the operating system, network, etc... So the devil is indeed in the details. It may be necessary to further tweak the product to make it work within some constraints (e.g., memory, disk space, processing capabilities of the machine, etc...) or to work with other components (e.g., Single Sign-On, smartcards, third-party products, etc...).
The installation is part of the deployment. The tweaking of the product is generally done as part of the need that exists for interoperability. Tweaking is also further done for the sake of performance, since the product is, by default, meant as a generic install upon a generic platform and it is generally required that parameters be tuned to really take advantage of all the performance offered by the specific platform upon which the product is installed.
This is thus why the Deployment, Performance and Interoperability (DPI) lab is there at Lotusphere and open for its attendees to visit. Once the core developers have provided a specific Lotus Software product, the DPI Lab people are the ones who are there -- as part of their regular functions at Lotus -- to help Lotus customers install products better, make them perform better and make them work with other products better. The lab has people with expertise on many different platforms (Windows, AIX, iOS, zOS, Linux, etc...) and for different Lotus products. They do this year-round as part of work engagements. So, during the Lotusphere week, they've come to the conference to share their knowledge and expertise.
All that's missing is you. See you there. ;o)
Frederic Dahm | 16 January 2008 12:42:26 AM ET | Montreal, QC | Comments (0) | Permanent Link
It wasn't so long ago that everyone in the lab was helping close the lab, having committed time and effort during that week to greeting people who had come, hearing their questions and doing their best to provide expertise and advise. Some like me, even took some of those questions outside of the lab and provided longer, more detailed answers by mail. As we closed the doors on the lab, the rest of the year lay ahead of us, where we would do roughly the same things as we did in the lab, except going from customer to customer and doing our best to help them fulfil their goals and objectives relating to their environment and their investment in Lotus (and connex) brand software.
Now, we are but a few days from another Lotusphere and come Sunday the 19th of January, the doors are going to open once again to lab members to setup the lab; the tables will be put up, the chairs placed next to the tables and some servers -- needed for some demos -- will be also setup. Then on the Monday morning, we'll all go back to the lab, open the doors, place our laptops on the tables and await once again our customers and welcome them as we did a year ago. As this is my third year in the lab, I look forward again to seeing Lotusphere attendees and hope to be as much of assistance as I was in the preceding years. So will my colleagues.
So bring yourself on over and come and say hi. It'll be really nice to see you. And if you've come to see us in the past years, it'll be nice to see you again. Truth be told, we never tire of the friendly faces of our customers. :o)
Frederic Dahm | 15 January 2008 11:43:39 PM ET | Montreal, QC | Comments (0) | Permanent Link
Rare Opportunity - Not to be Missed!!
Key Test Engineers on hand at the Deployment Performance and Interoperability Lab at the Dolphin Hotel in Europe 3/4 who can help you through some of your most challenging product areas: Test leading to Product Deployment! After learning about the new product features that you want to implement and deploy, you have to carefully plan for integration in your user installed base and production configuration environment.
The experts to help you through this process are available for an informative and insightful discussion, You will come away with new ideas and approaches to be more successful. Guaranteed!
Seriously ;),
The Lotus Connections System Test Team invite you to visit us and learn about IBM's approach to Test. We can share how we approach the many phases of Test at IBM including unit test, feature test, system test, performance test, globalization test and so on.
You can also learn of the key ingredients required to build a production readiness test plan, looking at all aspects of your IT challenge and structuring a test plan to address the rigorous requirements around production readiness.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 04:23:03 PM ET | | Comments (0) | Permanent Link
The Lotus Quickr System Test Team invite you to visit us and learn about IBM's approach to Test. We can share how we approach the many phases of Test at IBM including unit test, feature test, system test, performance test, globalization test and so on.
You can also learn of the key ingredients required to build a production readiness test plan, looking at all aspects of your IT challenge and structuring a test plan to address the rigorous requirements around production readiness.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 04:20:19 PM ET | | Comments (0) | Permanent Link
The Lotus Quickr System Test Team invite you to visit us and learn about reliability, load, stress and stability testing, as well as how to demonstrate that your infrastructure is able to meet the demands of your user groups.
Likewise, we can discuss product capabilities that maximize stability, and help you approach and plan deployment decisions in a way that maximizes value from your IT investment.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 04:17:42 PM ET | | Comments (0) | Permanent Link
The Lotus Connections System Test Team invite you to visit us and learn about reliability, load, stress and stability testing, as well as how to demonstrate that your infrastructure is able to meet the demands of your user groups.
Likewise, we can discuss product capabilities that maximize stability, and help you approach and plan deployment decisions in a way that maximizes value from your IT investment.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 04:15:21 PM ET | | Comments (0) | Permanent Link
The Lotus Connections System Test Team invite you to visit us and learn about the integration and interoperability capabilities that are supported between all of the Lotus Connections' services, including Profiles, Communities, Blogs, Dogear and Activities. Likewise, we can show you the integration capabilities that exist between the Lotus Connections services and other product in the Lotus Brand - such as Sametime, Notes & Domino.
Likewise, we can discuss your product deployment plans and help you approach and plan deployment decisions in a way that maximizes value from your IT investment, as well as sharing real System Test Data from testing and deployments that we have in IBM.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 02:42:11 PM ET | | Comments (0) | Permanent Link
The Lotus Quickr System Test Team invite you to visit us and learn about the integration and interoperability capabilities that are supported between Quickr and other Lotus products like Notes & Domino, Connectons, SameTime and Portal.
We can also show you integration capabilities with Microsoft Office and Windows explorer. Likewise, we can discuss your product deployment plans and help you approach and plan deployment decisions in a way that maximizes value from your IT investment, as well as sharing real System Test Data from testing and deployments that we have in IBM
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 02:22:45 PM ET | | Comments (0) | Permanent Link
How many times have you wondered about the complex configurations that we setup to simulate your production environments? Want to understand how we design and put these configurations together?
The Lotus Connections System Test Team invite you to visit us and talk to us about the Lotus Connections' configurations you would like to put in to production in 2008.
We can provide advice and consultation based on real data from IBM's Lotus System Test team We can also help you with recommendations for deployment topologies that will meet your specific needs in terms of scalability, reliability, fail-over, etc. We can also share internal IBM data from our own internal production environments.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 01:36:28 PM ET | | Comments (0) | Permanent Link
The Lotus Quickr System Test Team invite you to visit us and talk to us about the Lotus Quickr configurations you would like to put in to production in 2008.
We can provide advice and consultation based on real data from IBM's Lotus System Test team. We can also help you with recommendations for deployment topology's that will meet your specific needs in terms of scalability, reliability, fail-over, etc. We can also share internal IBM data from our own internal production environments.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 01:26:26 PM ET | | Comments (0) | Permanent Link
The Lotus Connections System Test Team invite you to visit us and talk about your configurations and experiences. Hopefully together we can share information that will improve our test environments and test configurations. We are happy to share our System Test approach used on Lotus Connections, and help explain the fundamentals underlining a good and solid System Test plan. And we'd like to hear about your approaches and challenges. We're hoping for a win-win from this dialog exchange with you.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 15 January 2008 01:19:39 PM ET | | Comments (0) | Permanent Link
Learn about IBM Lotus Domino Server.Load V8. From installation to configuration to running workloads, this article shows you how to quickly get started.
http://www.ibm.com/developerworks/lotus/library/domino8-serverload/index.html?S_TACT=105AGX13&S_CMPíU
Joe Malek | 15 January 2008 10:23:59 AM ET | Orlando, FL | Comments (0) | Permanent Link
We've been sharing our Lotus Connections deployment information with you - and can be found here:
Lotus Connections Deployment wiki
http://www-128.ibm.com/developerworks/wikis/display/lotusCollaborativeDocumentation/Home
The Lotus Connections team has been contributing to this Wiki, sharing their deployment experiences and configurations. Customers also are participating. Please review, learn from experiences and contribute your own!
However, while at Lotusphere, you can talk directly to the engineers responsible for this content. We've got the key engineers who contribute to this Wiki available to consult with you on your experience, plans and requirements. Representatives from Release Management (Carol) , System Test (Pat, Morten), Performance (Hunter) and Early Programs (Ray) are available. Come by the lab and talk with us for an in-depth, individual exchange.
Carol Zimmet | 13 January 2008 11:01:33 AM ET | | Comments (0) | Permanent Link
The Lotus Quickr System Test Team invite you to visit us and talk about your configurations and experiences. Hopefully together we can share information that will improve our test environments and test configurations. We are happy to share our System Test approach used on Lotus Quickr, and help explain the fundamentals underlining a good and solid System Test plan.
Patrick O'Sullivan is the System Test Architect and Morten Kristiansen is the Chief Tester. Just ask for Pat or Morten when you visit our lab.
Patrick O’Sullivan | 11 January 2008 12:05:11 PM ET | | Comments (0) | Permanent Link
Learn about the IBM Lotus Notes/ Domino V8 enhancements to the Server.Load tool and workloads. Three new workloads, N8Mail, N7Mail, and N8MailInit, demonstrate the typical tasks of a Lotus Notes user interacting with a Lotus Domino server.
For more, see: http://www.ibm.com/developerworks/lotus/library/notes8-workloads/
Joe Malek | 10 January 2008 04:58:24 PM ET | | Comments (0) | Permanent Link
As part of a trial by Advanced Technical Support (ATS) in the DPI Lab, we will are looking for several Customers who are willing to collect data about their production environments and then discuss/review it in the DPI Lab at Lotusphere. We will be taking the first three Customers who reply to this blog entry from each of the following Domino enterprise environments: System i, System p, System z/zOS, and System z/zLinux. If you are interested, please respond via email to mwojton@us.ibm.com before January 7th, 2008 indicating which one of the above platform you will submit data from. Thank you very much to those of you who have expressed interest. We have currently filled our available slots and will not be able to accept any more samples.
This trial will require you to collect data from 1 LPAR/OS image in your production environment along with the associated Domino servers within that LPAR. You will need to collect this data and have it received by IBM no later than Tuesday January 15th (one week before Lotusphere). You will need to collect statrep and native platform data (from outside of Domino). While three days is the minimum collection period needed, we would recommend 1 full week to include a sample of your weekend and off shift processing as well. Detailed instructions on how to collect this data will be sent out in the first week of January to the first three replies for each listed platform above.
Prior to Lotusphere, this data will be integrated and processed to build a web based graphical output that will be used for discussion with you at the DPI Lab. We will setup a 1 hour time where we can sit down in the DPI Lab and review this output with you and let you provide feedback about how we are presenting and using this data. You will receive a CD copy of the output.
As a starting point to begin a discussion of Performance Analysis and Capacity planning of a dynamic Domino production environment, here are a few items:
· Understand the difference between data that should be used for performance analysis and capacity planning. For example, CPU Busy is not a performance indicator but a capacity indicator. A CPU drop can be just as deadly to your server’s responsiveness as a CPU spike. These drops can be caused by a network, IO, memory, Domino or other bottlenecks that limits the work a server can perform. Typically, rates are capacity indicators while response times or cost per unit measurements are performance indicator. This means that performance data is not the same as capacity data and using one for the other purpose will lead to incorrect analysis and diagnostics.
· Not all problems are CPU related. How well do you monitor your IO infrastructure? What do you monitor? Are you aware that the IO device with the largest number of IOs quite often will not be the one causing you the greatest IO delays? While IO rates can tell you how much work a device is doing (capacity), it the response times of the IOs that will tell you how well it is doing that.
For example, a device that is doing ½ the IOs of another device, but has 3 times the response times will cause you a bottleneck before the device with twice the IO rate at 1/3 the response times. Devices can have differing response times because of workloads (more writes to log.nsf, mail.box(es), etc..) or different hardware (internal versus external, older generation versus newer generation drives, levels of caching, fragmentation, shared SAN resources being impacted by another DPAR or LPAR etc..) for the same IO rates. By using a combination of rates times response times you can calculate a device’s IO intensity to determine what devices is causing you the most delays.
Also, by tracking each Domino process’s CPU and IOs, you can monitor this close relationship as shown in the chart below. Deviations in this relationship (example shown below for 1 server task) can be an early indicator of either workload changes or performance problems well before your end users see them.
· A Domino workload is typically very dynamic. In order to be able to understand and interpret your Domino server’s usage of hardware resources, you must be able to relate these two together. For example, is a Router task that spikes and doubles its CPU usage a problem? If the cost per message delivered stays flat, then no, you had a mass mailing and this is a capacity issue, not a performance problem. However, if the cost per message doubles even though CPU usage stays flat, then you have a performance issue that can range from a defect to a networking problem or anywhere in between (but it is not a workload/capacity issue).
Examples of some of these cost that you can use to monitor your environment are: cost per active 15 minute user (shown below), cost per transaction, cost per message, cost per HTTP hit, cost per unique HTTP address
· In the large enterprise environment where you are running multiple LPARs and Domino servers on the same physical box and you have logical partitioning instead of physical partitioning, it is possible for one of these environments to impact the other. This is called ‘sympathy sickness’ where the DPAR/LPAR exhibiting the worse issue is merely the victim of some one else. Any shared resources can cause this same issue. For example, a large IO workload spike (backups running that saturate the SAN cache with sequential read ahead IOs) can impact a different Domino server’s performance that is doing random IO. In an enterprise environment, you must be able to see all LPARs, Domino servers and none-Domino workloads to fully interpret what is happening.
· How do you track the historical data for your environment? If the first time you are looking at your data is in the middle of a problem, then how do you know what is normal? Domino and your platform provide you with a large range of capacity and performance data. Being able to relate data that you captured from before a problem occurrence to data captured when the problem is occurring can significantly help you in identifying what is going on. The chart above shows you an example of the previous four years of history for the cost per active 15 minute user for a pair of Domino servers.
· Breaking out the cost of your various Domino operations. Do you know how much it is costing you to run clustering, agents, full text indexes? How about third party software such at anti-virus, BlackBerry, or backups?
For example, the majority of the cost in clustering is not in the cluster tasks (clrepl, cladmin, cldbdir) but in the server task. When clustering communicates with another server, it is performing a NRPC call. The only task in Domino that can bind to your network and receive these calls is the server task. Therefore, all of the cycles for clustering on the receiving side are in the server task (same is true for replications, etc…).
Different transactions cost different amounts of CPU to execute. The connections doing the most transactions may not be the ones driving the most CPU usage on your servers.
Mike Wojton | 20 December 2007 01:47:46 PM ET | | Comments (1) | Permanent Link
The results are now out, the feedback is positive so far. Let me know if you have any comments or suggestions.
Click here to download the presentation as PDF.
Ivan Dell’Era | 27 September 2007 10:26:53 PM ET | | Comments (0) | Permanent Link
