vBulletin Search Engine Optimization
| |||||||
| Register | FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| Hi there, Can anyone give me advice on what i need to do to set up the data repository for oracle enterprise? I'm afraid i am something of a novice to this and for various reasons have been given this. I have four databases at the moment and my tnsnames.ora is configured as i can use the visual studio with no problems. I just can't seem to get my head around this repository thing, can it be set up on a workstation or does it need to be a server? Any help would be appreciated..... I'm not familiar with most of this stuff. Thanks amani |
| |||
| The OEM repository is really just a special schema in any database. Like the RMAN repository, you can put it anywhere it's convenient. I prefer to create a 'DBAs only' instance on a totally separate machine and put the DBAs stuff there including OEM, RMAN, my version of workflow, etc. (Basic specs - 512M memory, 8G disk & processor speed > Intel Pentium 500 or equiv in a non-intel world .... my Sparc-266 Ultra5 limps a bit for this!) Remember, the big parts of OEM are a) Console, b) optional Management Serer, c) Repository and d) Intelligent [sic] Agents. There are 2 main ways to use OEM - 1) directly against the database/instance in question and 2) using the OMS as a centralized access point. Case 1 & 2 are not mutually exclusive. In case 1 ) the first time the console accesses the instance it will create a respoitory in that instance automatically. Thereafter you can do most of the management using this variant if you want, but there are limitations (IIRC, such as no jobs, no paging, etc.) In case 2) you decide which database & instance will host the respository for all the managed instances; you fire up the OMS and when you access the OMS from the console the repository will be created automatically. The OMS and the Agents will work together to collect all the relevant information into the centralized respository independant of how many consoles are connected to the OMS. Since this centralized repository will grow over time, and it contains all the job & event details, I really prefer to put this on it's own (properly administered) machine and instance available exclusively for the DBAs. This machine could be a workstation or a small server. Either way, if there are multiple DBAs this could become a heavily loaded machine. One suggestion might be start with an unused workstation and migrate to a server if load demands. HTH /Hans amani wrote: > Hi there, > > Can anyone give me advice on what i need to do to set up the data > repository for oracle enterprise? I'm afraid i am something of a > novice to this and for various reasons have been given this. > > I have four databases at the moment and my tnsnames.ora is configured > as i can use the visual studio with no problems. I just can't seem to > get my head around this repository thing, can it be set up on a > workstation or does it need to be a server? Any help would be > appreciated..... I'm not familiar with most of this stuff. > > Thanks > > amani |
| ||||
| "amani" <amanclark@yahoo.co.uk> wrote in message news:ddbedb82.0306260708.a3419d7@posting.google.co m... > Hi there, > > Can anyone give me advice on what i need to do to set up the data > repository for oracle enterprise? I'm afraid i am something of a > novice to this and for various reasons have been given this. > > I have four databases at the moment and my tnsnames.ora is configured > as i can use the visual studio with no problems. I just can't seem to > get my head around this repository thing, can it be set up on a > workstation or does it need to be a server? Any help would be > appreciated..... I'm not familiar with most of this stuff. > > Thanks > > amani What's not to understand? The repository for Enterprise Manager is just a bunch of tables stored in a database somewhere, in which details of what jobs/events have been scheduled are stored. If you run the Enterprise Manager configuration assistant, it will actually propose to create you a working database into which this repository can be dropped (or, in plain English, within which the necessary tables can be created). Otherwise, you can persuade the EMCA that you've already created an appropriate database (use Database Configuration Assistant and select any of the templates) and have the repository created in that. So I'm afraid I don't understand the confusion. The idea of the repository is that it is in a database to which you have ready access, so it's probably more appropriate to create it on a workstation than on a server... unless you have half a dozen DBAs who all want to access EM simultaneously. As for how you get started, and thus eliminate the confusion you're suffering, fire up EMCA on a workstation to which you have easy access and which has had the complete Oracle stuff (not just the Client) installed. Select to create a new database, call it OEMREP, and let te wizard do its stuff. Since you are creating a bunch of tables, someone needs to own them (the 'repository owner' or the OEM schema owner). Pick a username such as REPOWNER... you'll never actually log in as this user any way: it's merely an account to own the repository tables. Then you need to start the Management Server process (or, Management Server service on NT/W2K/WXP): on Unix, you do oemctl start and on windows you click the "play" button for the Management Server service (hint: mentioning an O/S would help whenever you post). With the Server started, you can log in to it. Username = SYSMAN, password=OEM_TEMP. With that done, click Navigate- Discover Nodes, type in the name of your server that has the real database you want to manage, and you're off (or should be: the Agent needs to be running on the node you specify, and the LISTENER.ORA needs to have a static SID_LIST, rather than rely on just automatic instance registration). ~QM |