JHT_Seattle
Hot Shot
Hot Shot

AppVolumes 3.0 - does anyone have it working?

I can't get through the basic config wizard.  Using the same credentials/info as a working 2.10.0 environment, the 3.0 appliance just reports connection failures.  Logged into the appliance console and it can ping everything successfully.  Anyone else out there having any luck?

0 Kudos
12 Replies
dray11
Contributor
Contributor

No.  Similar experience to yours.  Although I don't think the 2.10 creds are relevant.  I installed the appliance using the ova.  I can log on to it at the console (root/123), change the password, check networking... all good.  But cannot get any response from http://<applianceip>/horizonadmin.

0 Kudos
cmbwml1
Enthusiast
Enthusiast

I'm stuck as well.

First issue was with static IP.  The appliance booted but got stuck waiting for DHCP so I added registration to DHCP and rebooted appliance.  When it came up there was a problem with the AD configure button.  I redeployed the OVA and made sure to configure DHCP with new MAC before booting.  That worked and the config button allowed me to add AD and Admin group.  I'm stuck at the vROPS config.  Documentation says that vROPS is optional even thought the app volumes config wizard shows it as required.  I'm going to move on the vCenter registration and see if the config finishes.

What step are you stuck on?  You may need to add the appropriate certificates to the app volumes appliance cacert.pem file.  I added our internal AD CA, public CA root and intermediate certs, etc.

0 Kudos
JHT_Seattle
Hot Shot
Hot Shot

I've tried selecting the option to accept any cert, and I've tried disabling cert checks in the appliance, neither has resulted in me being able to establish a connection to either my vCenter (Windows-based, not the VCSA) or my vROps instance.

I opened a case for it, I'll reply back here if/when anything improves.  Awaiting 3.0.1...

0 Kudos
chulerico
Enthusiast
Enthusiast

you might need to have certs disable to test

/etc/wemi/utils/disable_ssl_validation.sh

the setup and usage is definitely different than 2.10, no computer assignment, no longer has option to assign/de-assign stack on the fly, file share requirement, do see some of these options better suited for stability and easier to integrated different deployments (example multiple vcenter, etc).

vROps is optional

when you install agent, take note that it install vrops, uem, and appv components.

good luck.

Sam

0 Kudos
JHT_Seattle
Hot Shot
Hot Shot

I did note that I disabled cert checks... made no difference.

I do understand that vROps is optional, but if I'm opting to connect to my existing vROps instance it should be able to connect, and it isn't.  vCenter is more important, obviously, and it can't connect to that either.  I'm pretty sure it's some issue in the appliance (perhaps the firewall), but I'm not about to go digging through this.  It appears that enough people have issues with this first release that I'm opening the case and letting VMware dig through it so that by the time we get a release where migration is an option the appliance will be stable and a viable option for upgrade.  For 3.x to succeed it needs to provide some additional value over what we can do with the 2.10.0 release.  Currently, there is none which is why everyone is advising to stay on the 2.10.0 release.

0 Kudos
dray11
Contributor
Contributor

I think I was having the same problem as you.  I got it working finally... but not sure what if anything I did.  I reinstalled the Appliance three times.  The third time it worked.  Fortunately the install is pretty quick.

The first time I installed it with the old J# vSphere console.  Felt like it went ok but like you, it seemed like the networking did not work. 

Decided to try again in case I accidentally did something wrong.  This time I used the web console.  Seemed to go about the same.  Again, networking did not work.  This time I was determined to figure it out.  I am no Linux expert but I can get around.  I figured out that I could only ping my local network from the appliance.  Nothing could ping the appliance... not even machines on the local VLAN.  I checked everything... it seemed anyway... must have missed something.  Certainly confirmed all the eth0 interface settings.

So... I finally decided to try again.  Deleted the appliance and installed again... from the web console.  This time it worked.   ???

I can't make any sense of this.  Probably something about me or my environment... but that makes this feel a lot more like a finicky tech preview and less like a GA.

0 Kudos
JHT_Seattle
Hot Shot
Hot Shot

Okay, after playing with the HTML5 vCenter Fling today, revisited my AppVol 3.0 appliance.  I created an IP Pool for the subnet I deployed it to and associated it with the port group the appliance is on.  Tried to register vCenter again, SUCCESS first try.  Tried vROps, still no good.  Maybe it's just randomness?

0 Kudos
sappomannoz
Hot Shot
Hot Shot

Hello,

for me the configurations steps all work, but I'm stuck when scanning the file share for new apps. I get this in the logs and the no new app appears!

[2016-03-29 07:59:34 UTC] P2103R14230 ERROR Unable to detect an ESX hostname for "#<Xms::Vcenter:0x00000007a6f098>"

[2016-03-29 07:59:34 UTC] P2103R14230 INFO Completed 400 Bad Request in 296.3ms (Views: 0.2ms | ActiveRecord: 1.3ms)

[2016-03-29 07:59:34 UTC] P2103R14230 INFO

Cristiano

0 Kudos
sappomannoz
Hot Shot
Hot Shot

It works now. Datastore prefix is case sensitive, but to fix it I had to redeploy the ova, so maybe there where also other errors.

0 Kudos
dray11
Contributor
Contributor

I feel like I am on a different problem track and should stop making noise here... but wanted to followup on my previous comments for anyone else who comes along.

I figured out that the appliance firewall blocks ICMP.  Therefore do not expect it to respond to pings unless you disable the firewall (/etc/wemi/utils/disable_firewall.sh).  It is fine to have the firewall on unless you need ICMP it for testing (I think).

The console gave me confusing results because it would work in Chrome but not in IE... only I didn't notice the pattern for a while.  Turns out I needed TLS 1.1 enabled in IE (IE 10... why wasn't it enabled already?  I dunno.)

Anyway, I suspect if I had not focused on pinging and had tried Chrome first I would have connected to my first install.

Now I have successfully AppCapture'd, Assign'ed and Agent'ed to get a simple app on my test client.  Modest success.  Now can do real testing.

Good luck.

0 Kudos
Hayes_marc
Contributor
Contributor

Has anyone been able to remove apps from appvols 3??

0 Kudos
eeg3
Commander
Commander

Unfortunately, there is no way to remove apps from inventory built-in; release notes indicate to contact VMware Support to remove them.

Blog: http://blog.eeg3.net
0 Kudos