There are often a wide variety of reasons why someone may want to keep only the latest n revisions of content in their content server. These reasons range all over from keeping disk usage in check to a form of security/retention management. Whatever your reason the process is pretty straight forward. I am going to show you how to setup archiver to take out all but the latest X revisions. In a later post we can talk about various ways to enact that archive on a scheduled basis.
So, simply put, the secret to this is the Revision Rank (dRevRank) field. A lot of times in the past I've tried looking at dRevLabel, dID, dDocID, but nope, the special field you want is Revision Rank. Use the following steps to set up your archive:
- Log onto your content server as a user with administration rights
- Under "Administration" select "Admin Applets" and then fire up the Archiver applet
- From the "Edit" menu select "Add"
- Give your archive some kind of meaningful name
- Provide a description, perhaps include something about the expected lifetime of this archive?
- Click "OK"
- Select the new archive in the archive list and then click on the "Export Data" tab
- From the Export Query section select "Edit"
- Set the following field defintions
- field to "Revision Rank"
- Set the Operator to "Is Greater Than"
- Set the value to "4"
- Click "Add", add and then "OK"
- Under "Actions" click on "Export"
Great, now lets discuss some of these points. The Revision Rank field is zero based (meaning it starts at zero) and that zero represents the most recent revision of the content in question. The second most recent revision is denoted by a dRevRank of 1, etc. For example, if you wanted to delete everything but the 5 latest revisions you should set dRevRank to 4. If you wanted to only keep the most recent revsion set dRevRank to 0.
Upon selecting "Export" you will be prompted deleting the revisions from content server during the export process. If you place a check mark in the check box the content will be REMOVED from content server. The content will continue to exist in the archive you just exported as long as you don't crack that open and delete it in there as well.
In his keynote, Larry had several interesting things to say. First, he announced a new product from Oracle. This is, in fact, their first ever hardware product. It is called The Exadata Storage Server. I wonder what Network Appliances (a major sponsor of OpenWorld) thinks of this? The Exadata Storage Server can be standalone or placed in a grid of these things. He also talked about a new connectivity model between the database and the storage server, named InfiniBand. This allows more data, faster, from the disk system to the database.
Hey look, another new product! This one is called The Oracle Database Machine. Good Lord, it has 64 Intel Cores for database processing and 112 cores for storage processing and gigabytes and gigabytes of RAM. And it holds some ridiculous number of Terabytes of data, something like 168.
Get More Info
While I’m waiting on Larry for his presentation I wanted to recap one of the session from earlier today that I have not got around to posting about yet. Yeah, I’m a slacker. The presentation was an overview of JDeveloper and the Application Development Framework.
The presenter was very good, and he moved very quickly from topic to topic. The points of the presentation were very good, but the demo was better. It was awesome. I’m getting ahead of myself. Oops. I am going to try to sum this up: ADF allows Oracle themselves to utilize developers of various degrees of expertise to quickly and efficiently build useable, robust, efficient applications.
Ok, let’s talk about the demo. That’s why we were all there anyway right?! In just a little under 25 minutes the presenter (Shay? I’m probably spelling that wrong) created a page that paged through data, displayed several master detail scenarios and allowed updating. The whole thing looked nice and worked quickly. A lot of AJAX was built in out of the gate. Graphing was built in. All in all, it was very, very cool.
I have to do some thinking on how this works with ECM. I hope they will soon have prefabricated ADF components to work with the Content Server, or even WCM. How does the Open WCM concept fit into this? How? Why? So many questions, so few answers....
Overwhelmingly I hear people refer to the leader of the Oracle Enterprise by just "Larry". Does he know we're all on a first name basis with him? I wonder what he thinks of that? It's kind of comical, and maybe representative of several social, psychological paradigms. Are we attempting to socially elevate ourselves by referring to a person of power and wealth on a first name basis? Is it something else? I sure don't know. But it did strike me as interesting.
Some of you may have used or even still have an excel spreadsheet which can be used to update metadata. This spreadsheet was originally provided by the former content server creator, Stellent. I am actually unaware of the name or names of the original author or authors and as such have not credited them here. Leave your name to take credit! Some people love it. Some people hate it. Either way, it has been useful to me.
The original style sheet allowed users to specify additional custom columns named after their custom metadata. It offered two buttons. The first was “submit query” which invoked the user defined query string located to the left of the button. The second button “update” would cycle through each of the result records from the query and update the metadata based on the values in the spreadsheet.
I have updated the spreadsheet in two ways. I have added a login button which simply causes a prompt for user credentials. The importance of this button is the reduction in steps needed to update secured content. With the original spreadsheet the user first had to search on public content only and invoke an update which then prompted for credentials after which the user could conduct another search which would then have the credentials to access secured content. With the login button we can now skip this cycle. The second change is a new worksheet which contains configuration variables for the searches. Currently there are three configuration variables including ResultCount, SortField and SortOrder.
Additionally I often get asked how one can use this spreadsheet to update metadata only check-in’s. The trick is an additional column named createPrimaryMetaFile. This column should be set to TRUE if that record represents a metadata only check in.
This file and all updates to it are presented here as-is with no support implied or otherwise. As always, use at your own risk.
References:
remoteMetadataUpdater_20080924.xls (194 KB)
Session-wise, I would like to hit these today:
| Time |
Room |
Description |
| 9:00 AM |
Marriott Golden Gate C1 |
Basics of SOA Deployment for Enterprises |
| 11:30 AM |
Marriott NOB Hill AB |
Using Oracle SOA Suite and Oracle BPEL Process Manager to Integrate and Extend |
| 1:00 PM |
Marriott Salon 5 |
Reshaping Your Business with Web 2.0 |
| 5:00 PM |
Marriott NOB Hill CD |
Optimizing User Engagement, Using Oracle Real-Time Decisions |
In between all of these sessions I am hoping to finally get down to the demo grounds today. The overall goal for the day is to learn as much as possible about SOA Suite.
This session opened with a recap of active content versus historical content. Active content was defined as documents, web content, images, etc. that are being actively developed. Historical content was content that was not accessed, forwarded or replied to in the last thirty days. It was noted that do to various litigation/liability issues historical content needs to be online and discoverable instead of collecting cobwebs on tape in a remote location.
We learned not to forget about the traditional reasons for storage of historical content via an archiving solution such as reduced, consolidated storage and the most common scenario: disaster recovery. These things still exist, but the legality and liability points are much stronger drivers and are immediate, real, substantial threats. Alongside these reasons for needing archiving it is important to note the difference between an online solution versus near-line or offline systems.
The basic idea described was to archive content out of a source system with the OPTION of leaving behind a stub or links and then take the core content and place this into the Universal Online Archive. A tool named High Volume Importer was described for loading content to the system. The system (somewhat because of the underlying requirement of Oracle 11g Database) supports optional compression and de-duplication of content. An example of an 8 core machine was used as a baseline in which they outlined a test case of accomplishing the ingestion of 15 million emails a day.
A set of services were outlined to work with or on top of UOA. The first was the currently shipping Oracle Email Archive Service. In addition a service named File Server Archive Service got some discussion, and it sounded pretty neat. The intention of this service was to archive files from a windows network share or file server. The Oracle Archiving Interface is a set of services that comprise the High Volume Importer and a basic API for developers work with so that you can develop your own services on top of UOA. The coming soon list of services included: Application Archive Service (transactional data, etc.), SharePoint Archive Service, ECM Archive Services (includes non-Oracle solutions such as Documentum), Image Archiving, and finally Desktop Archive Service.
It would appear that ZL Technologies has a death grip on the market for services built upon UOA at this point. It sounds like they have several of these services built out and even a few more. All of their services were noted as being developed against UOA via the Oracle Archiving Interface.
It was unfortunate, but this session seemed more like a long advertisement for partner based services built on top of UOA. This information is important to get out, but there was little discussion UOA itself or any form of deep, let alone middle, dive into UOA itself. Other topics were covered, and overall this was very informative, I just wasn't the right target audience. :)
They did give a really killer example of using this with email archiving though and in the examle the cost of storage of email was reduced by 98%. This seemed to not take into account the COST of buying UOA (oh, and it requires a license of Oracle Enterprise Edition Database), but was neat anyway.
Bob Fraser, a Principal Product Manager for Enterprise 2.0, presented a session named No REST for the Weary: REST APIs and the Activity Stream. This centered around working within WebCenter. Of course, we start off with an overview of what Web/Enterprise 2.0 is. For the 3 people at the conference that don’t know.
Interestingly, this session was less about REST itself and more specifically about where and how it can be used in WebCenter, which was great. He also talked about security with REST/WebCenter so he satisfied that basic need.
A small byproduct of the demo was some discussion of curl, a command line URL invocation application. Available out of the box on most new Mac and Linux systems this application can be handy for automation scenarios and development. It is also available on Windows.
This was really my first look at WebCenter Interaction, a product in the WebCenter Suite. It is neat, but I am again surprised at the level of overlap with other products. Certainly if you wanted get an application with features X, Y and Z you would have several options from within Oracle alone.
References:
cURL
Presentation [PDF] (2.33 MB)
I figured I have seen the Multisite Web Content Management presentation seventy-four times now so I opted instead to go see a session about WebCenter/WebLogic/UCM. Just prior to that we stopped by the Fusion Middleware Lounge to visit with Brian Dirking and see who was out and about already. Then it was off to the session. Official title: Deploying a Web-Oriented Architecture with WebCenter/WebLogic.
A quote about WOA (Web Oriented Architecture) from Dion Hinchcliffe’s Blog:
"In other words, the Web model provides a single, open, and unified information architecture that is consistent, easily consumed, extremely scalable, securable, very reusable, resilient, and highly federated." Wow, that sounds very sales-ish to me.
There was, once again, more advertising about REST. This has been mentioned again and again. Remember, REST is not really a new technology in the sense of age or implementation. It is simply a different way of looking at web services. Make the time to learn about it though, fairly easy once you get into it. Having said that though, I do not know as though I’ve seen anyone implement a REST style call system for Content server…anyone else? Something like:
http://myserver/idc/DOC_INFO/
or
http://myserver/idc/DOC_INFO_BY_NAME/
Would this be useful, would anyone care? Forever and ever Stellent, oh sorry, I meant Fusion ECM Content Server, already had URL based accessibility, though I wouldn’t call it strictly REST.
Anyway, back to the session. They showed a little bit of AquaLogic Ensemble which was used to pull some information off Facebook and then integrated that with a CRM solution. Kind of neat.
There was also a fair amount of discussion about using Dojo JavaScript Library with WebLogic and a thing called Disc, some other framework. They also promoted JSON a little bit. The WebLogic demo was very Web 2.0 oriented demonstrating changing things in real time on the client side of the interaction.
In the end, the demos were somewhat hurried and the message seemed to be more about REST/JSON than about WebLogic. While WebCenter was in the title I do not believe it was ever mentioned again. Maybe that was the brief look at Ensemble in the beginning?
References:
WebCenter
WebLogic Portal
Dojo
Example WebLogic Portal Site
Several sessions look interesting today, especially at 5:00 p.m. Like six of them. Sheesh.
9:00 a.m. Deploying a Web-Oriented Architecture with WebCenter/WebLogic
9:00 a.m. Multisite Web Content Management
11:30 a.m. REST API's with Activity Stream
1:00 p.m. Oracle Ensemble and New Integration Patterns
5:00 p.m. Universal Online Archive
5:00 p.m. Oracle Total Recall Hands-on Lab
What to do, what to do. I have until 9 or so to figure out how to split myself to go to multiple places at the same time. Don't hold your breath.
|
Copyright © 2009 Jason Stortz. All rights reserved.
|
|