Does anyone have any recommendations about how they setup the Documentum component of PPM in an HA configuration? Our instance of Documentum currently resides on a Windows 2003 box and it's Oracle database sits on the database server (Windows 2003) that is shared with PPM.
The Oracle server will be clustered with another server using Oracle Data Guard, and the PPM instance will be on redundant servers with a front-end hardware load balancer.
first of all. what do you exactly mean with Documentum component of PPM? Do you mean the default document handling of PPM or the additional installation of EMC Documentum for Integration with PPM. I'm sorry but i also do not understand the word "HA".
Sascha: I am trying to make the EMC Documentum component (which exists on its own server, separate from PPM) highly available and/or redundant. The plan is to have 1 PPM server in each of the 2 data centers with a load balancer in front of them, 1 database server behind each PPM server that are kept in sync with Oracle Data Guard, and then 1 Documentum server in each data center as well.
I'm not familiar enough with EMC Documentum and I wanted to know how others were making the Documentum server redundant in situations like this. Are they using packages like Microsoft Clustering Services or Veritas Cluster Server, or are they using something native to EMC Documentum to provide the HA and/or failover functionality?
Attached is a diagram to help illustrate what I'm trying to accomplish.
your illustrations looks good. We are running nearly the same architecure but not so highly available, like you wants to have.
From my point of view, it is not necessary to configure a scenario that the documentum content server become high available. We are running an old PPM 7.5 SP7 cluster (two physical windows server, each handles two ppm nodes) with documentum 5.3 SP6 integration (windows too). Furthermore we configured a documentum index server to enable full text search for our users.
Because we configured a very complex demand management with alot of automatic request creation and mandatory documents, we attached hundred of documents a day. Actually, the users attached nearly one million documents. This "huge" amount of data does not have any impact on the system, presupposed you have enough file space :)
First of all, you should ask you the following questions:
1. How many requests will be created montly?
2. How many attachment field did you configure?
3. How many of these fields are mandatory?
4. What will be the average size of the document?
5. Do you want to use Full-Text Search?
This will help you to get a better feeling for your decision. I didn't configure a high available scenario for documentum because this server will not become the bottle-neck by high load. We made several performance test and normally the ppm database server has to manage most of the data.
However, if you plan to use documentum, you should know some base configurations at the repository. Otherwise some internal jobs from documentum will spam the instance :D