Support on VC 6.0 + ESXi 6.0 was tested many times for a long time in the QA.
Which VC exact version are you using?
Please try the following steps:
1. Un-register the plugin (using the PowerCLI ps1 script)
2. Log-out -> log-in in vSphre web client
3. Clear your web browser cookies / history etc.
4. Restart the vSphere web client service (Virgo server) on your VC
5. Register the plugin (using the PowerCLI ps1 script) -> remember NOT to press Enter until you perform the log-in into your vSphere web client (that completes the registration)
6. Once you see the SIO icon in the vSphere web client, you can press Enter in the PowerCLI session (that will stop the embedded tomcat)
1. Was this a VC 5.5 and you upgraded it to 6.0? or was it a fresh install / deployment?
2. Did you try the steps I wrote in comment number 5?
3. You verified that there's no connectivity issues / Firewall / wrong VC credentials were used?
4. When using the PowerCLI script, remember NOT to press Enter until you perform the re-login into your vSphere web client (that actually completes the registration). Pressing Enter before that stops the embedded web-server and will fail the registration.
I've had this issue on the two VCs I've tried to install the plugin on (I still have it on the second).
I'm using a fresh install of the latest VC appliance (6), 1.32 plugin, and latest version of PowerCLI. On the first system, after 10 - 20 attempts, it finally registered. On the current system I've lost count of the number of attempts (I've been trying for over 7 hours now).
I've followed the instructions above to the letter, and tried other variants suggested on these forums, but none have succeeded. Very infuriating.
The service shows up under objects, but is listed as stopped / unknown. I'd appreciate any help at all.
Did you perform all steps to clean up your system + cleared the browser's cache/cookies?
Here are the steps again - please clean your system (and the Virgo logs) and repeat the registration procedure and then send us all logs + files (see relevant paths below).
How to clean-up your Env and perform clean install (while using previously defined networks):
IMPORTANT: Unmap and delete any ScaleIO volumes that you have in your system
If needed: Unregister your ScaleIO system (from within the plugin)
If needed: Delete all your ScaleIO SVMs
Unregister plugin (old build)
Stop vSphere web client service (VC Linux: service vsphere-client stop)
Delete the content of the folder (VC 5.5 plugin location): C:\ProgramData\VMware\vSphere Web Client\vc-packages\vsphere-client-serenity
Delete the content of the folder, or folder itself (VC 5.5): "C:\Windows\System32\config\systemprofile\AppData\Roaming\VMware\scaleio"
Clean the Virgo logs folder: C:\ProgramData\VMware\vSphere Web Client\serviceability\logs
Start the vSphere web client service (VC Linux: service vsphere-client start)
Register the new plugin (new build)
If needed: Remove the SDC module params + vib itself from the ESXs by connecting via SSH to each ESX and run these 2 commands:
~ # esxcli system module parameters set -m scini -p ""
Thanks for the response. Can I get some clarification? If I am ready your steps correctly, I have to uninstall the entire previous ScaleIO deployment and start from scratch? I'm in a catch 22 situation if that's the case, as I can't delete and of the volumes as they have many VMs on them with no option to move them to other storage temporarily.
Some background info on the situation - the original deployment went fairly smoothly. ScaleIO was up and running for a couple of weeks without issue. The original VC became corrupted and no backups were available. The only option was to build a new VC and re-add the PD. The existing PD is functioning correctly in the background, so it seems like a tall order just to install a plugin in a new VC.
The environment is not live yet, but I'm loathed to dismantle SIO to such an extent that several weeks of work will be undone over what should be a simple plugin!
Carl-Adam_Lilje
32 Posts
1008
0
Posted June 17th, 2015 10:00
Yes, I installed it in a test environment just this week, worked like a charm.