On Jan 13, 9:26 pm, Corey Osman <[email protected]> wrote: > Hi, > > I am somewhat new to puppet and I wanted to show a demo that automated the > install of oracle to my DBA in hopes to entice him to start automating his > procedures. > > I have linked my manifest code below for review. In the code I have > attempted to automate a complete oracle install by following the instructions > provided by my DBA. Its very rough code as I am not an expert yet. > > http://pastebin.com/Qw39iP7z > > The code assumes that the oracle pre-install checklist has been completed via > another class called oraclebase. I am using oracle generated response files > to automate the actual oracle install. However since oracle doesn't come > inside an rpm I don't know of a good way to check if oracle is installed. > Using du is probably the worst way since after an install it will take too > long to determine if the size is bigger than 5 (line 16). > > I thought about using NFS to install which would change my cwd to > /net/nfsserver/oracle_install/database which would save bandwidth but there > may be times when NFS is unavailable or slow WAN links.
It was the very first module I wrote back in 2009, so no guarantees that it is the best or prettiest approach, but the set of classes linked to below can install either the oracle server or client. http://pastebin.com/nVnrkf2J I was not aware of the oracle-validated rpm, I'll have to check it out. I wouldn't trust it for the oracle user and groups, since those IDs should be consistent across systems and rpm won't ensure that. For some of the other setup (e.g. sysctl) I worry that it would clobber existing configuration. > How can I tell if oracle is install or any program when its install without > rpm? Check for a file it creates, like lsnrctl. For example, from my oracle::install_base class. exec { run_installer: cwd => "$tmp", creates => "$oracle_home/bin/$installer_creates", command => "su - oracle -c \"$tmp/$pkg/runInstaller - ignoreSysPrereqs -waitforcompletion -silent -responseFile $tmp/ $response_file &> $tmp/runInstaller.log\"", require => [ Class["oracle::packages"], Exec["extract_installer"], File["response"], File["inventory"]]; } > How can I tell when the untar command is finished? This strikes me as the wrong way of thinking about things. Anything that can't happen before the untar should require the exec that performs the untar. > What happens when it takes a while to transfer 1.8GB file? As far as I know while one step in your manifest is being done, puppet won't do anything else. As others have said, be wary of using a file resource for this. > Does the "user => oracle" give me the environment of the user like su - > oracle would? No. > Is there a provider for patching oracle? Is there a patch type? No. This is something the DBAs where I work handle themselves using oracle-specific tools. > Is it bad to have lots of exec statements? It's messy, but sometimes there's no an obviously better way, or the amount of work required to do it more cleanly can't be justified. I'm not sure that I see the advantage of your rpm scheme, for instance, since you're clearly not going to upgrade oracle by putting newer rpms in your repository and running yum. If you do go the route of installing oracle and copying files, rather than running the installer on each server, be sure not to miss files, like /etc/oratab, that will end up outside of your oracle base directory. Also, be sure to discuss this thoroughly with your DBAs and test to make sure it doesn't break any of their tools; I wonder if there might be identifiers in the oracle inventory catalog that are supposed to be globally unique, for example. All the best, Brian Pitts -- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
