Wednesday Jan 30, 2013

Managing Log Files

I've seen a couple of questions recently about log files in Ops Center - where do we keep them, and what kind of automatic management do we do for them?

There are quite a few log files that are used in Ops Center. The Logs and Directories appendix in the Feature Reference Guide provides the locations of these log files, organized by category.

Also, some of the logs can be managed automatically. The appendix also explains how to automatically manage the numbers of some log files.

Wednesday Jan 23, 2013

Workflow: Monitor and Manage Incidents

The Monitor and Manage Incidents workflow is one of the mostly new workflow documents. It explains how to monitor assets and deal with incidents.

 It's built around this image:

The left side has the prerequisites; you have to set up hardware, libraries, and operating systems before you can get into all aspects of monitoring and incident management.

Once you've gone through the prereqs, you can get into monitoring and incident management. The Understanding OS Performance and Capacity how-to explains how to use OS analytics, and the Tuning Monitoring Rules and Policies how-to explains how to fine-tune the monitoring for all of your assets. Managing Incidents and Using Service Requests explain how to deal with incidents and file service requests for them if necessary. Finally, you're given a pointer over to the Feature Reference Guide, which goes into more detail about all of these features.

Thursday Jan 17, 2013

Recreating a deleted OVM Server guest

I got a question from an Ops Center user:

"I created an OVM Server (aka LDOM) guest through the Ops Center UI, but it got deleted by accident. It's been recreated by hand with the same configuration, vdsdev, vnic, vdisk names, everything. Is there a way to 'import' it into Ops Center, so that it can be migrateable within a server pool, without reprovisioning it again?"

As long as the control domain is healthy, you can do this.

First, you run a deployment plan to re-create just the OVM Server Container, using all of the same properties (disk, network, etc.) as when it was first created. A normal guest deployment uses a deployment plan to set up the container and an OS profile to complete the guest; in this case you're just doing the first half of that process. Make sure that the guest doesn't exist and that no vdiskdev has been created. Once you've created the container, it will be in the 'installing' state, ready for provisioning.

Next, you do the following on the control domain, in order to boot the OS in the new container:

# ldm set-var boot-device=disk:a <guest name>

# ldm set-var boot-command=boot <guest name>

# ldm stop <guest name>

# ldm start <guest name>

Finally, you have to unconfigure and reconfigure the Agent in the guest. You can do this manually, or using the Add Asset option in the UI. Once you've done that, the OS will be running in the new container, and will be fully migrateable.

Thursday Jan 10, 2013

How To: Enable Single Root I/O Virtualization in Exalogic Elastic Cloud

The Oracle Exalogic Elastic Cloud X2-2 solution includes a set of compute nodes connected through QDR InfiniBand IO Fabric. In order to virtualize the I/O for these nodes, the Single Root Input Output Virtualization (SR-IOV) has to be turned on. However, ILOM and BIOS updates will reset the settings, which prevents virtualization, and causes some headaches if you're trying to get that virtualization working quickly.

So, we put together a How-To that explains how to enable SR-IOV. You can do it through the Ops Center UI or from the command line; the how-to walks you through both ways of doing it, and gives you a pointer over to the Exalogic section of the Feature Reference guide for more info.

Wednesday Jan 02, 2013

Workflow: Deploy Oracle VM Server for SPARC

The Deploy Oracle VM Server workflow is one of the workflow documents that we've added. It explains how to get Oracle VM Servers for SPARC up and running.

It's pretty straightforward, and it's built around this image:

The left side shows the prerequisites. You have to have Ops Center installed (most options in Ops Center have this as a prerequisite), deploy libraries, deploy hardware, and deploy operating systems.

Once you've done that, you can begin your deployment. You configure and deploy the VM servers, then create your server pools, and finally configure and install your logical domains. Once you've done that, you're directed to the Operate Logical Domains workflow, which picks up where this workflow leaves off.


This blog discusses issues encountered in Ops Center and highlights the ways in which the documentation can help you


« January 2013 »