HP CloudSystem 8.1 Part 2
So, I've been banging my head on the deployment of this product at home but have since gotten some hands on experience in work to start getting my head around concepts and today I've retried setting up my Lab from scratch to rebuild CS8.1 and get it to work. I consolidated my SSDs to two largish ones to help with space requirements but this wasn't a show stopper as much as a nice to have. I was using an LSI card but in JBOD mode, found this wasn't great in terms of performance so now each SSD is allocated to a RAID 0 array albeit with one drive and things have improved with it's SSD caching ability.
None of that is really relevant or interesting (!) to what I'll describe here - the biggest piece of the puzzle came when my I broke the CAT-5e cable I had running downstairs to my Lab. I had strange issues with my netgear prosafe switch so after some research ordered a Cisco SG300-10 but subsequently discovered the cable issue, the switch was actually fine! The reason I went with the Cisco is to get layer 3 capability plus it's a much more powerful unit albeit with a web front end and not a full cisco IOS. I saved a fair bit by buying via Newegg but had to get a shipping forwarder to send it onto Ireland, didn't get hit with Duty so it worked out very well. The unit is €330 here, I got it for $172 + $72 forwarding charge (It's a heavy item) as an example. The Netgear power supply worked with it perfectly which was a bonus! There is a trick to getting it in layer 3 mode via a serial cable but they supply it so don't worry, I've it hooked into my vCenter VM by adding a virtual serial interface to that VM in ESXi!
Now, I was able to spend some time recently understanding the CS8.1 networking and after I redid it from scratch I deployed all the VMs without the previous HP-OO password issue. I found that the Enterprise appliance ignores the template values in terms of vCPU & RAM, it demanded 8 vCPU and 20GB Ram for itself, the cheek! I powered it off and applied 4 vCPU and 8GB of Ram and restarted it to see if this would work. My Lab only has 4 physical cores after all!
The Networking piece is described below to help you plan your own deployments. I think it's key to have a good Switch/Router and not just rely on ESXi for cloud stuff in particular.
| vLAN ID | Subnet | Gateway | Port Group |
| 1 | 192.168.10.0/24 | 192.168.10.254 | vLAN_Cloud_DC_Mgmt |
| 50 | 192.168.11.0/24 | N/A | vLAN50_Cloud_Mgmt |
| 51 | 192.168.12.0/24 | 192.168.12.254 | vLAN51_Cloud_CAN |
| 55 | 192.168.13.0/24 | 192.168.13.254 | vLAN4095_Cloud_Data_Trunk |
| 56 | 192.168.14.0/24 | 192.168.14.254 | vLAN4095_Cloud_Data_Trunk |
| 57 | 192.168.15.0/24 | 192.168.15.254 | vLAN4095_Cloud_Data_Trunk |
| 58 | 192.168.16.0/24 | 192.168.16.254 | vLAN4095_Cloud_Data_Trunk |
| 59 | 192.168.17.0/24 | 192.168.17.254 | vLAN4095_Cloud_Data_Trunk |
| N/A | 192.168.1.0/24 | 192.168.1.1 | vLAN_Cloud_External |
My Default Network for all my existing VMs was using subnet 192.168.10.0/24 so I left that as the native vLAN. I pretty much used the same settings as before with a few changes as shown below:
I updated my Windows PC Hosts file to make sure I could communicate. Now, I get as far as the Enterprise Appliance but it's refusing to start up and my Foundation Appliance is going crazy, 100% cpu! Looks like I'll need a Hardware Upgrade to deploy Enterprise at this rate! At least I can play with Foundation and it's portal until I figure out what's happening. I would suggest that if your lab is like mine this might be something you'll have to do on a work server not at home. I'll post any updates if I manage to get this working. At least we've no errors as shown below during the Foundation Appliance Deployment:
start>csstart gui --start-browser --auto-accept-cert
Web server starting.serving on http://127.0.0.1:5000
Config file - passed basic tests, moving to advanced tests.
Config file - passed advanced tests.
Config file - passed basic tests, moving to advanced tests.
Config file - passed advanced tests.
Creating new base appliance.
Warning: Found 4 cores on the hypvervisor. Decreasing core request from 8 for ap
pliance.
Appliance (ca1) successfully reconfigured
Booting the appliance.
This step could take between 5 and 20 minutes to complete.
Elapsed time (minutes): 2
Finished.
The CloudSystem controller is being started.
This step could take between 10 and 20 minutes to complete.
Elapsed time (minutes): 6
Finished.
Waiting for the CloudSystem services to finish starting.
This step could take between 5 and 15 minutes to complete.
Elapsed time (minutes): 0 Complete.
Configured appliance EULA and support access.
Applying the first time setup network selections.
Using ssl cert:
-----BEGIN CERTIFICATE-----
XXXXXXXXXXXXXXXXXXXXXXXXXXXXX
-----END CERTIFICATE-----
VM started successfully.
Open browser to https://192.168.10.70/Update: Power Cycling the Foundation Appliance resolves the issue. You can then go in and choose to "uninstall" CloudSystem Enterprise, it just seems to reset the integration until you win the lottery and can afford a massive home server!
Update #2: After adding in the Compute Cluster you get to choose which vSwitch the Cloud Data Trunk gets created on. I noted that the proxy is actually a clone of the base appliance and like Enterprise it doesn't take it's hardware settings from the template so prepare for a 4 x vCPU and 16GB Ram VM in your environment. I've since downgraded it to 2 x vCPU and 8 GB Ram to see how it plays.
