I’ve been doing a bit of work over the past number of weeks on the adapters for vCenter Operations (vC OPs) with my old pal Paudie. We are working on vCenter Operations 5.8 and using a vSphere 5.5U1 environment. Since we have a Brocade Fibre Channel switch and an EMC VNX array in our lab, I wanted to get the Management Pack for Storage Devices (MPSD) and the Brocade SAN Analytics Management Pack deployed, and see what information we could glean from those extension packs. When we completed the configuration, we were able to go into the vC OPs customs view and see details like the following Brocade – Health Overview and Storage Components Heatmap:
Caution: We spent a lot of time trying to figure out why the MPSD adapter would not connect to the CIMOM service on Brocade’s Network Advisor. This boiled down to networking/DNS configuration issues. The MPSD release notes for vC OPs describe the issue. As they say, I should have RTFM. Anyhow, here are the steps we went through to get this setup going. I’m afraid it is rather long, but hopefully you will find the information in here useful.
Step 1: Setup the Brocade switch to generate SNMP traps
This is probably going to vary on a switch by switch basis. Since our switch was a DS5000 from Brocade, these are the steps that are needed to generate an SNMP trap when a change occurs on the switch. The Brocade Network Advisor SAN User Manual has further details.
Track changes status: OFF
Track changes generate SNMP-TRAP: NO
DS_5000B:admin> trackchangesset 1,1
Track changes status: ON
Track changes generate SNMP-TRAP: YES
Step 2: Register VASA Provider for the VNX array
We burned a lot of cycles here as the procedure to register the VASA provider for the VNX has changed a few times, and there are a number of blog posts out there which detail how to set up a VASA provider for older versions of flare code. We are running flare code 32 on our VNX 5500, and found that the correct procedure was documented here – Configuring VASA with VNX blog post. Once VASA was registered (and note that there is no need for Solutions Enabler or SMI or anything like that now), we could see the array capabilities surfaced up in vCenter:
Step 3: Install & Configure Brocade Network Advisor
The next step is to install a piece of software called the Brocade Network Advisor (BNA). This contains a CIMOM server and SMI Agent, and both must be configured during the installation:
You will now need to add any Brocade fibre channel switches which connect your ESXi hosts to the storage array, in our case, to the EMC VNX. These switches have to support Brocade Network Advisor. The switches are added via the Network Advisor – Product Status and Traffic. In the first screen shot, no switches are visible. You need to click on the Discover link and add FC switch details.
In the next screen shot, the discovery is complete and the FC switch is now visible. This process needs to be repeated for all switches in your fabric which connect the ESXi hosts to the storage arrays:
Step 4 – Install the vCOPs PAKs
The PAKs can be found on the VMware Solutions Exchange web site. The Brocade SAN Analytics Management Pack and the MPSD – Management Pack for Storage Devices can both be found on that site.We decided to install them both. Note that these are completely independent PAKs – there is no reliance on one to the other. However both display different information about our fabric and storage, so we decided to add both.
These packs must be installed via the admin interface of the vC OPs UI. In the Update tab, browse to the PAK and click Update. You will get a warning to say that you will not be able to revert to the previous version of vCenter Operations Manager, so be sure to have a backup to revert to if something goes wrong. The status tab will inform you if the new adapter from the PAK is running after installation. In the screenshot below, you can see that a new Adapter called STORAGE_DEVICES is now running successfully:
Step 5 – Configuring the Adapters and Credentials
There are the final steps., and the appropriate documentation should be referenced for the correct and exact steps.
Basically, you need to do a “Describe” operation, and then create new adapters and new credentials for the add-ons to vC Ops. This process is well described in the documentation, but to do a “Describe”, you must login to the vC Ops custom UI (https://vcops-ip-address-or-hostname/vcops-custom), navigate to ADMIN > SUPPORT > INFO, and from Adapters Info click on the icon for “Describe”, select Yes and then OK. This is not the easiest icon to find, but it is located in the upper right hand corner of the Adapters window. This describe operation finds the adapter files, gathers information about the abilities of the adapter, and updates the user interface with information about the adapter.
Once that is completed, you can now go ahead and create adapters for both the Brocade Switch and the Storage Devices. You will need credentials for the BNA to hand, as well as the IP address of the BNA host and the ports used by CIMOM (24600), the SMI Agent (5989) and SNMP (162). These are available via the BNA if you wish to double-check. Again, the documentation for the Brocade SAN Analytics Management Pack and the MPSD – Management Pack for Storage Devices can both be found on the VMware Solutions Exchange site. The creation of the Adapters also has a “Test” button to verify that everything you added is correct.
As I mentioned in the introduction, we had some issues with the MPSD communicating to the BNA CIMOM. This turned out to be nothing more than DNS/network related, and the work around in the release notes solved the issue for us. The Brocade SAN Analytics just worked out of the box, and no issues were encountered.
Step 6 – Build Customer vC OPs dashboards using this new data
At this point, you are now in a position to build your own custom dashboards using this new data from the Brocade switch and the EMC VNX storage array. A number of default dashboards are provided, however more can be built if you so wish.
I am not going to cover this here – but I may return to it at a future time. However there are a number of bloggers out there who have step-by-step instructions on how to do this. A quick google for “vC Ops custom dashboards” should lead you to some.
Key Takeaways/Lessons Learnt
The steps which cost us most time could have been quicker if we had done the following:
- Ensure that all hosts participating in this configuration components have DNS fully configured and that both forward and reverse lookups are working.
- Ensure that you are implementing a VASA provider that is correct for the storage array and firmware version on that array.