Hello. In this video, I'm going to be demonstrating how you can perform a manual RN back up through the use of an MD A. So I'll be covering um an overview of a manual initiated arm and backup along with the creation of an arm and script for that backup. So with N MD A um installed on the oracle server um and a client resource on the network or backup server defined for the oracle server. We want to back up in this case, Mac or one, We can uh begin the process of creating an R man and launching a manual backup from that oracle server. So typically that process is initiated by the uh the DB as the oracle user.
So I will become the oracle user in my environment. And um uh a manual arm and backup is basically an execution of valid arm and script with a whole bunch of NSR variables, N variables uh specific to the N MD A backup. So here's what that arm and script that I have here looks like for this uh manual backup. So as you can see, as you can see here, this is my arm man script and it starts off with a connect string. Um Although I don't need to have the connect string inside my arm man script. Um In, in this case, I do have it. The reason why I don't need to have it is because since it's a manual arm and backup, there's nothing preventing me from launching Arman and passing the connect um the user name, password in my in, in the connect string manually, something like this and then calling my arm and script. So as you can see here, I provided the uh connect string outside of the R manuscript.
So I don't need to have this inside the R manuscript, but it's just an option. Um Next, we go into the actual um uh main component of the Arman scripts, which is the channel allocation and the rest of the backup options that we need to specify. So after the run uh clause in the open parentheses, I'm allocating a couple of channels here, which is basically a degree of parallelism for the oracle backup. And uh I have, I have to have SPT tape defined so that oracle knows that it's going to send that back up stream to a media management library that's been previously linked with oracle. This was done as part of the N MD A installation. Uh Next, it is important to specify some NSR variables in the R manuscript and I'm doing so using the send command.
So these networker environment variables. NASR EV. Here, I'm defining who is in this case, who is the network or backup server? So my network or backup server name is called Mac win one. So that's my network or backup server and the network or client that I want to send that data to uh is MAC AA one. So this is actually my oracle server called MAC or one. It's very important that you specify these two variables in your R man because uh well, the backup server is pretty obvious you want to define where you want to send your data to. Uh But then the nicer client variable is very important, especially for cluster environments where you want to define the client file index that's gonna uh contain the uh the backup that you're sending um that you're generating here through this manual process.
So if you don't define uh NSR client, let's say in a cluster environment uh doing restore if you doing a restore from a node in your cluster, other than the one that did the backup, you could potentially not be able to find the backup pieces generated. So it's always very important to have NSR client to find um in your arm man. Uh There's a whole list of uh network or variables that are mentioned in the N MD A administrators guide that has a whole slew of um options you can pass along in your arm and backup. Uh Here, I I, I'm specifying the, the media pool that I want to send my data to. So on the backup server, I've created an or a pool and a network or media pool called oracle. And um I do not want my manual backup to go to the default pool. Hence, I'm specifying NAS a data volume pool equals the name of the pool that I want to direct my, my data stream to um next in this backup uh script. I'm I'm saying that I'm doing a backup uh level full and I'm backing up the entire database which um will include the uh current control file.
And in this case, the SP file as well. Uh And I'm specifying my false per set to four, which describes the uh uh arm and multiplexing that will take place along with uh giving uh the name of the format that I want the back of pieces to be generated as. So basically all these backup pieces that are going to come from the oracle server are going to be save sets within the network or server using this format. In this case, percent D stands for the database, instance, name and percent capital U guarantees a uniqueness. Some alpha numeric number will be generated for the backup piece uh with that done.
I then release my two channels and that's the contents of my manuscript. So why don't we go ahead and run this now? So the way you can run it. Uh Well, there's several ways of, of doing so because I've specified my connect string here inside my arm and script, I can just simply do this. So as you, as you can see here, it found my arm man script and I have N MD A 1.6 installed. So oracle is able to load this up and start the backup because it saw that um we have allocated a channel that's type SPT tape and the backup has completed here now. So here are all my backup pieces for the various data files that were saved. In this case, these three files are contained within this backup piece.
And if I were to come over to um if I were to come over to the network or backup server, you can see that those backup pieces should not be inside uh uh the, the backup server for um uh for, for backups uh that are initiated from the oracle client. You would normally see the level here as manual, but you would define the actual level of the backup within your R man script. So if you want to do an incremental backup versus a, a level four, you would have to write the appropriate R man script to state that.
Thanks for watching. I hope you found this video helpful and stay tuned for more videos to come.
