|« MP Studio v4.1 Now Available||Join the SharePoint and System Center Live Meeting »|
Review: SCI live meeting on managing Sharepoint with System CenterDPM
This was another very interesting meeting.
Here is a short step by step of the contents:
Leslie Kits - Program Manager on System Center - started by giving us the agenda of course and some introductory words. This gives a framework to give an idea of the importance to actually manage Sharepoint farms as they become more and more prevalent and complex.
SharePoint 2007 was the fastest growing application in Microsoft history from a user perspective with more than a milion users. First it was mainly used as a departmental tool for sharing documents. Later more and more team collaboration sites sprang up and versioning was leveraged. More and more sensitive data gets stored on Sharepoint sites. In most cases the IT department was not really in control because of lack of knowledge of all the inner workings of the product and mostly power users were creating and working with the sites. Problem is when something goes wrong. Sensitive and critical data can get lost.
If a SharePoint farm is maintained and managed in a good way this is proving to very cost effective. IT needs tooling to centrally manage products and the System Center line is most suited for the job of bringing things together. The most used seems to be automated backup as a cost saver on labor. Automated patch deployment is a possible big cost saver as well.
System Center brings us proactive monitoring and reporting, managing of capability and provisioning, rapid data backup, optimal configuration of SharePoint, auto deployment of patches, automated compliance framework, manage virtual and physical environments, manage globally spread components, consistent user interface across multiple applications.
Next on was Durgesh Nandan - a program manager on the Office Internet Platform. Talking on the SCOM management pack for SharePoint 2010.
The focus of the new MP was to improve design monitoring with the larger architecture of SharePoint itself. Monitoring for SharePoint 2010 is about logical monitoring. This links from the physical layer to the logical layer to determine the effects on the logical components when a physical component stops working.
A big improvement is SharePoint Health Analyzer rules (SPHA). These show up in the central monitor and system center monitor. It looks like it works like a best practises analyzer equivalent.
The new management pack is a unified MP. First it was distributed across multiple MP's. This would create a lot of MP's for each component. Now it is offered as one MP. It discovers what is running. This will auto adjust to the number of services running. It shows the physical topology view and the logical topology. For instance - one web app can be supported by multiple instances. So if one goes down the service as such is still running. The SLA objectives will need to reflect this and the right targeting for this objective can be set.
All the rules surface in the SPME rules state view. This helps sort and it will show the Sharepoint health analyzer rules that give errors or warnings in an easy way.
Durgesh showed a graph that shows how monitoring is improved. The number of classes, monitors and discoveries have grown a lot. for example from 10 classes to 133. This improves the granularity of monitoring a lot.
Also the new SPHA rules and some technet knowledge has been added. Technet knowledge was included in the MP at first. Now all knowledge is located on technet, so it can be updated constantly. For 150 cases of problems a technet article is linked already. This will help a lot as well.
Number of general rules and reports have decreased. This was because a number of them were redundant because they were already included in other management packs, such as SQL. This reduced noise a lot. The SP 2010 management pack assumes that the SQL MP is installed. Some SQL related monitors are in the MP but are disabled by default.
There are also less reports because of double entries with the SQL MP's etc.
Jason Buffington - our data protection hero was next on DPM 2010.
The Release Candidate is due next week and the RTM is expected in the spring of 2010. So go and get the RC when it gets out next week
So specifically to the SharePoint parts:
SP 2003 does not have VSS writer. So SQL is backed up and STSADM can dump the SP config and can get backed up.
For SP2007 and windows sharepoint services 3.0 the VSS writer is used. As long as agents are installed on all servers in farm you can point DPM to one of the front-ends in the farm. DPM will find the components in the farm, like sql content servers, enterprise search, farm config database, system state of frontends. Using other products was difficult because of the distributed nature of SP. A lot of components placed everywhere. DPM is designed for this.
If topology changes in the farm - just add new server + deploy dpm agent. Next time a scheduled backup is run the farm will notify the change to DPM. The DPM admin needs to acknowledge by clicking in an OK box for this to become active in order to acknowledge that the admin knows that additional space on dpm server is needed for this addition.
Demo - Looks like a demo of the RC ;-)
A short demo was given on protecting the SharePoint farm. This is easy as long as all servers involved have a DPM agent installed. Selecting the Sharepoint farm on one of the frontend servers will make DPM find the rest automatically.
A feature is highlighted. Self healing. Automatically perform consistency check if the replica is inconsistent. This rules!
On the topic of recovery. Well for SP 2003 the separate parts need to be restored. For SP 2007 restoring requires a recovery farm. Normally a single server. Take the content db and bring the data back to production.
For the new SP 2010 no recovery farm is needed anymore. Restoring single items directly from DPM to SP 2010 is now possible.
Leslie Kitz came back for more:
sharepoint virtualization best practises. In short most front end services, such as web services can be virtualized very well and performance and high availability can be improved by scaling out. Backend functions can be virtualized, but the heavier loads will probably go towards more physical components. Of course this can be combined.
DPM 2010 for sharepoint white paper will be coming up.
300 level webcast is coming up as well, but expected in March.
Working on customer case studies.
For customer evidence contact Leslie Kitz (dpm 2010, SCOM mp's), customer evidence, case study or technet edge videos and blogs, references.
The next meeting will be on March 9 on Essentials and DPM (I bet we are going to see Jason again!).
So on a personal note from myself:
It is no secret that I just love the System Center line of products and have been working with these a lot. I believe the added possibilities in the SharePoint 2010 management pack for SCOM will make the product much better manageable and give some control back to IT personnel that is less specialized in SharePoint. I have been working with DPM 2010 Beta for some time now and am very happy with its workings and self healing possibilities. And the coming additional features in the release candidate will bring even more enthusiasm. More to come soon...