DBMS service dot create service and call it my serve network name my serve and that’s being defined only in this container I then go to another one connect sis / horrible at JW to assist EBA and do that you get slightly older fax it because if I select service name from DBA services hmm I will think to myself oops select name from DBA services I got that error trying to gauge something cause crying to create my serve I’m told already exists no it doesn’t it does but in another container. So this is one of those cases where you get a bit of a where you guys know normally I sure you can understand it’s necessary you can’t have the same service name created in two containers because the listener wouldn’t know which contains to connect you to, so it’s a bit of an anomaly though that you’re told something exists you can’t actually see but of course in the root container.
You can see the truth because in the root container we can select the name corn ID from CDV from cdb services and we can see it will exist but he won’t have been started yet we’ll do that in both times yeah now you get the idea yet once it started of course then you could work it out yeah, so there’s the truth of where it is yet once it started the listener will also tell you and then instance parameters there is one SP foil for the internal SED be only because there’s one instance but some parameters you can adjust locally per container some you can adjust per container some you can also just sub/container and there are some bugs as well right with the important books they’re just annoying bugs so if you look at v dollar parameter is pd be modifiable so cilix name mm is pd be more reliable from v dollar parameter order by 21 so I’ve got 381 parameters and the heck of a lot of them can be changed within a PDB and a lot of them can’t you can usually work it out.
If you think of the ones that cannot be modified than a PDB well anything to do with a memory you can’t possibly modify within a PDF yeah because to administer the instance you couldn’t have the DBA in one container adjusting yeah the size of your buffer cache because there’s anyone buffer cache you spice all the sessions no matter what container they’re connected to clearly you can’t change the DB name in a portable container so anything to do with the roots the control file anything to do with undo if you look at your undo settings, for example, you can’t change that within a container and that makes perfect sense, but there are some very interesting ones indeed that you can change within container look, for example, parallel query these you sat in the root things like parallel max service you can figure that in the root container but you’ll find within your pluggable containers you can change these and this is interesting.
So you can set up parallelism in the roots but control how it’s used per container a lot of the optimizer ones you can change within a container even for instance optimizer features enable so you can have different pausing going on in different containers very interesting indeed although you can tune them differently using these and if you look at say redo all this stuff or false and here’s a bug own here we are true and what the heck log archive guess one do you think you can change that’s an applicable container yeah alter system set log archive test 1 equals location equals C cone / TMP and you got a jelly yeah car lots, of course, you can’t and that’s what I mean when I say there are a few bugs in this clearly that’s ridiculous I very much doubt you can do job queue processes haven’t actually tried that what if you can I definitely maybe you could perhaps you could have a pool of job queue processes that was different in each container well I never you can that’s a nice one.
So you can have different job queue processes than per container or so it would seem but it depends on the nature of the parameter know whether the variable within a container or the scope is the whole CD by me oh yes of course some can be set to be ignored that I suspect may well be one of them and then the final topic of the converse of the chapter is our man and the redo log is, of course, managed in the root so you put the whole database into archive log mode I’m not going to demonstrate all of this because you’ll do it in the exercise so you put the database into archive log mode in the root because of course there’s only one stream of redo now you can’t have one database in archive log one container in archive log motion of the container an option archive log mode that will be impossible from the roots you can do everything almond has been enhanced so it understands the environments.
I will show you some syntax in a moment you can back up the entire container database orders one container because a container is nothing more than satyr tablespaces and of course you can back up individual tablespaces what is very nice though is you can do incomplete recovery of an individual container you can do point in time recovery of a container, so this means that you’ve got a CD to be with 20 pluggable containers you can shut down one of those pluggable complainers and do a restore and recover until so you can take one container back in time very nice indeed you can do that or of course the entire CD be can be recovered but point in time recovery of one container is really nice especially as the rest of the cdb remains online you do that in the exercise there’s some syntax within the pdb is what we’re always used to from the root however you’ve got commands like this and you’ve got restore pluggable database PDB and that can be a point in time recovery flesh back.
If you enable database flashback which I usually do that’s only done the cdb level I can’t think of a technical reason for that I can see no reason why you shouldn’t be able to flash back one container but certainly in the current release you can’t flash back is the whole CD be so I’ll take a look through the exercise fat so you can see some of the sort of syntax that I want us to do it’s what I want us to do if I can oops need to go our exercise worksheet for chapter 4 we assume we’re going to work with sort of shut down tablespaces users and roles backup restore flashback right we assume at the moment you’ve got see DBA with two PDP’s I think you’re all at that stage and some of your gone beyond that make sure things are working registered with the listener fine we know how to do all this there’s some syntax just moving around and question 3 is various ways of starting up and shutting down and there’s the creasing the trigger to automate startup question five experiments with tablespaces.
Now and we’re here we’re setting different permanence and deep default tablespaces differently for the different containers and you should see what’s going on their question 6 is undone I’m expecting some of this to fail and I need you to think about why you’ll also maybe find one or two things maybe you’ll find something succeeds that you think shouldn’t succeed well try again put some thought into that their things don’t work it’s deliberate it’s the same with common and local users a lot of what we’re doing here syntax for common in local users see if you can work out what you can do and what you can’t do see if you can work it out the same thing with roles have a go various bits of syntax here some will work some will not work try and get it straight in your head hmm so what will work and what will not work have a few goes yeah before you run the command think yourself it’s not going to work or not the reason.
I think this is really important I said on the OCP exams there was a huge amount on these topics and basically, you have to be able to look at any of these commands and think, and the question will be you run this what will happen a will succeed we will fail with this message see you will fail with that message and so on and you have to work it out without actually being able to run it these were difficult then some services creating service is starting services see what you’ve got then question 10 backup and restore. So we’re taking right the way through the cycle here backup and restore of pretty much everything we could think of shutting down the database is put into archive log mode do a full backup and then some experiments that syntax complete recovery of just one container and note that throughout all of this we’re doing complete recovery restored recover.