Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Mailing Lists -> Oracle-L -> RE: OEM justification -oem Poll

RE: OEM justification -oem Poll

From: Michael Fontana <>
Date: Wed, 28 Mar 2007 17:34:35 -0400
Message-ID: <>

Well done, Matt. Even if slightly biased, I think you did a good summarization on the top of using GC and other tools for monitoring.  

I like to think of monitoring as an auditing function, in a sense, and I question whether or not Oracle is going to necessarily expose it's own poor housekeeping and software failures when reporting on their own databases. I've already seen where they somehow will neglect to do so if it's suits their interests.  

I work in a large shop offering enterprise hosting to a myriad of customers. Absolutely none of them are clamoring for GC. They want things as simple as possible. It is probably an easy install in a lab, but try getting it to run in the real world where servers are in multiple physical locations seperated by firewalls and VPNs. That is one of the major problems I encountered - problem is, when GC can't report/install on a client, or when the client can't be "discovered" from the GC host, there are practically NO TOOLS available from Oracle to determine the problem. Even when you open an SR and it is figured out, the process is not repeatable. I could go on and on about the limits of GC, but I do agree that the free pieces are probably worth looking into, especially if you've just moved into a shop and want a good picture of the environment, and what needs fixing. But for day-to-day monitoring of important systems with reliable information, I would not even consider this product mature enough to be considered....      

[] On Behalf Of Matthew Zito Sent: Monday, March 26, 2007 1:30 PM
Subject: RE: OEM justification -oem Poll

Well, speaking as a thoroughly biased software vendor whose product is often compared to certain pieces of Grid Control, I have spent a fair bit of time looking at Grid Control and working with customers who use Grid Control, here's my .02:

I've spoken to many customers who say that Oracle pushed Grid Control heavily into their environment. Among people who have tried to implement it, experiences have ranged from "couldn't get it to work, oracle flew an engineer from corporate in, they spent a month, they couldn't get it to work" to "worked great out of the box". It seems, again, simply from the people I've talked to, that the diagnostic and tuning packs work pretty well. The provisioning and configuration management packs don't work at all (at least, never met anyone who has successfully gotten them to work). Even in the 10gR3 Grid control, provisioning and patching have huge swaths of issues.

In general, I think the big test should be - can you set up grid control in your lab, get it working, and get all of the packs you're going to spend money on working? Does it meet your needs? Are you better off picking best of breed technologies for different areas?

Again, I'm biased - we make a automation, provisioning, patching, and configuration management product for database environments. But none of this is my opinion, its just based on what I've seen people doing in different environments.


Matthew Zito
Chief Scientist
GridApp Systems
P: 646-452-4090

-----Original Message-----
From: on behalf of Spears, Brian
Sent: Mon 3/26/2007 11:18 AM
Subject: RE: OEM justification -oem Poll

I started the OEM poll and I will give a summary when I get a break in

But in response to your question..

- OEM is great for proactivity
- OEM is great for scalable monitoring
- OEM is great for Rapid tuning and problem detection (the newest guy in
the dba group can do it)
- OEM is great for helping with historical data growth
- OEM has a Waits Graph on system performance.(when u learn how to
use.You can immediately spot issues)
- OEM is great for looking at impacts of historical
trends...cpu,disk,memory We have hundreds of databases so it kicks butt in helping us be proactive... We could script most stuff.. But then you create something that is homegrown and not industry supported. Quite a few people have written some kick butt stuff and can manage pretty well without OEM for the basic stuff. But when it comes to it.. It wont match oem for scalability..if it does.. You are just trying to create your own OEM. OEM is quite extensible... It enables all kinds of time saving options... Examples: We set up oem accounts and oracle accounts in our databases and let others manage various repetitive duties like unlock account etc. Nothing to install.. Give them the link to point to and they are off. We have too much project load so we gave a limited access (one database with read only on their tables) to OEM to look at their own crappy sql and fix it..(when you are starving for time...this is a real breather) We have been using OEM for years and everyday we get new ideas of how OEM can save us time.. Just put in OEM jobs in the scheduler to automate bounceing/start/stop all dbs per server with hand off pages/emails to various groups ( In larger companies ...this can be a nightmare... Unix guys wont let you hook the startups in the startup directories ) The OEM scheduler is quite good ... And I see the technology improving in quality each release and that oracle is putting effort to make this a critical management tool....FINALLY! I find OEM like the Rabbit hole.. Mentioned in Matrix... You just never know when you reach the end of functionality.. I find a new screen everytime I go in... ( ya it suchs a bit on the user interface... Its not self educating or intuitive but one can manage) It goes on and on... You can monitor the app/web servers as well as the oracle database and know where the problem is coming from... Without wasting time... There is some great features like comparing hardware setups to check for differences.. Cloning dbs, patching dbs' with hook into Oracle metalink do down load patches etc. Could it be that they are they copying Microsoft auto patching (he he)?... I can see down the road this can create incredible functionality and time saving. Bottom line... Does your company need to be that on top of issues? We get new databases going in all the time so it really helps with problems of change. All this stuff takes some spin to do the benefit/cost proposal but Personally after I was forced away from most of my scripting... I have come to respect and value it (although that took a few years...and still some minor reservations). If your company is not that big on proactivity and they pay great... Drop me a line! Part of why we got into it was the that company wanted to reduce the risk of all this dependancy of dba scripting and have an industry common frame as well as vender supported. Where do I get my check from Oracle? Brian -----Original Message----- From: [] On Behalf Of Sent: Monday, March 26, 2007 8:33 AM To: Subject: OEM justification I creating a justification for OEM. (Should I?) The company used to license the packs for EM console, but stopped a year ago. I want them to license OEM Grid. For what I guess is $40,000? per 4 processor server? It would monitor 5 servers, 2 of them development, (maybe more... especially if it were to be extended). ADDR, Accounts for various activities... e.g. Refreshing Dev or Test databases from OEM himself without DBA :) (cloning). Managers, operations, can see real time graphs of how the database is doing. Jobs: RMAN backups, Tuning, and on and on. What is the impact of Not having OEM. The databases will humm along like normal for a couple more years... but? Just Listing all the benefits can be over the top -- they would also have to understand... and overcome the hurtle posed by the fact that we are doing 'fine' now -- aren't we? Any feedback? -- -- --
Received on Wed Mar 28 2007 - 16:34:35 CDT

Original text of this message