Jump to content


Photo

PVS 7.6 HA not working

Started by Alexander Kuchelbacher , 26 January 2015 - 03:50 PM
18 replies to this topic

Alexander Kuchelbacher Members

Alexander Kuchelbacher
  • 22 posts

Posted 26 January 2015 - 03:50 PM

Hi!

 

I have a HyperV 2012 R2 enviroment.

I stream over PVS7.6 an XENAPP6.5 Windows 2008 R2 Server.

I use for Booting a CD which includes both PVS-Servers.

I stream in RAM with overfloating to Clients Harddisc. This works wihtout any problem.

I use also 2 NICs, one Legacy for Booting and systehetic where Traffic is failovered.

The replication status of the virtual disks is fine.

Each PVS has it's own harddisc where the images are stored.

When I stop the streamingservice on one of the server or on the other, there is no failover happening.

LoadBalancing is activated on the vdisk.

 

Any idea?



Nathan Joseph Members

Nathan Joseph
  • 131 posts

Posted 26 January 2015 - 04:12 PM

Is the vDisk you are testing with in private image mode, or are you booting your testing target device from a testing or maintenance version? HA should not work with an image that is in read-write mode.

 

Are the timestamps of the vDisks on each PVS server identical? If the servers don't view the vDisks as identical HA will not take place.



Alexander Kuchelbacher Members

Alexander Kuchelbacher
  • 22 posts

Posted 26 January 2015 - 04:15 PM

the vdisk is in standard mode, as described replication status is fine. Time stamp is of course little bit different, because i needed to copy it from one pvs to the other one



Nathan Joseph Members

Nathan Joseph
  • 131 posts

Posted 26 January 2015 - 04:39 PM

I don't recall which of the three timestamps PVS examines, if it is just "Date modified" or also the Access date, so I'm not sure if that is OK.

 

You could try copying the file with robocopy and using the /DCOPY:T option which should preserve all of the timestamps, or the /COPYALL option that should preserve all file information and timestamps.



Andrew Edwards Members

Andrew Edwards
  • 33 posts

Posted 26 January 2015 - 04:45 PM

I'm having the exact same issue. If I try to restart the service through the console of both the local PVS server im connected too and the other PVS server, I get the 'Failed' message too. I'm sure its linked somehow but at the moment I cant figure out what. 

 

Im a little different in the fact im using XS6.2 and Xenapp 7.5 but that shouldnt matter as its the PVS HA streaming technology that isnt working. 



Andrew Edwards Members

Andrew Edwards
  • 33 posts

Posted 26 January 2015 - 04:53 PM

Is your PVS server licensed? Mine isnt and its the only thing I can think of. 



Alexander Kuchelbacher Members

Alexander Kuchelbacher
  • 22 posts

Posted 26 January 2015 - 09:58 PM

I have tested it with XenDesktop Platinum License, still same error.



Martyn Carey Members

Martyn Carey
  • 20 posts

Posted 26 January 2015 - 10:32 PM

In the vDisk versions < Replication do you have a tick in the versions on each server?

 

Do you have multiple NIC's in the PVS server?



Alexander Kuchelbacher Members

Alexander Kuchelbacher
  • 22 posts

Posted 26 January 2015 - 10:35 PM

Hi i do not have a tick in Version,

I only have one NIC on PVS.

Firewall is disabled.



Alexander Kuchelbacher Members
  • #10

Alexander Kuchelbacher
  • 22 posts

Posted 27 January 2015 - 12:36 AM

Just shut down one of the PVS. All target devices ran on the other PVS, but I needed to reboot them.

I copied all vhd and pvp Files from the running Server with robocopy and /COPYALL Switch.

i started the Soap and stream Server after copying.

than I tried to rebalance the devices.

This already did not work.

 

Any other idea?

 

Many thanks,

 

Alex



Andrew Edwards Members
  • #11

Andrew Edwards
  • 33 posts

Posted 27 January 2015 - 10:20 AM

Alex,

 

I'm the same as you,  only 1 NIC in the PVS server. 

 

Can you restart the services by right clicking a server > stream service > restart? I just want to see if you're getting the same results as me as I think it may be linked.

 

EDIT: The reason I couldn't restart the service was because the service account I setup wasn't being used to start the soap service so that's now fixed. Still having a problem with this HA stuff though. 


Edited by Andrew Edwards, 27 January 2015 - 12:49 PM.


Alexander Kuchelbacher Members
  • #12

Alexander Kuchelbacher
  • 22 posts

Posted 27 January 2015 - 01:05 PM

Hi . Service stops proper. As i Wrote before, Even load balancing ist mit working. But the Target devices are Able To boot from both pvs. I Shut down pvs2 and rebootet all Target devices. They Boot from pvs1. Also the other way around. But no ha or load balancing

Andrew Edwards Members
  • #13

Andrew Edwards
  • 33 posts

Posted 27 January 2015 - 01:31 PM

Thats exactly the same as me. I've just tried and I can boot from both PVS servers fine, but if I stop the stream service on one server while machines are connected to it, the machines just halt instead of being streamed from the second PVS server.

 

I've gone through all sharing and security permissions today on the stores, the services and the database itself but still cant figure out why its not working. 

 

Suggestions anyone?

 

Thanks,



Alexander Kuchelbacher Members
  • #14

Alexander Kuchelbacher
  • 22 posts

Posted 27 January 2015 - 10:48 PM

great i am not the only one.

I have a lot of experience in PVS, but here I stuck. Logfiles from cdf Monitor does not Show any suspicous Thing.

Strange.

 

Any other Suggestion?



Andrew Edwards Members
  • #15

Andrew Edwards
  • 33 posts

Posted 28 January 2015 - 10:10 AM

Same, I've never had this issue before. I cant image Citrix releasing an issue without testing the HA especially when 7.6 is supposed to contain lots of previous fixes.

 

Is there anyone else out there using 7.6 in production or non production that they can test? I literally don't know where else I can go with this.

 

Thanks, 

 

P.S I've changed my client to PXE boot now and have configured the Bootsrap with both servers but am still having the same issue.



Alexander Kuchelbacher Members
  • #16

Alexander Kuchelbacher
  • 22 posts

Posted 29 January 2015 - 09:50 AM

Hi, just fixed it with Support.

What a bad error.

It was the licensing. Today after one week I saw license Errors in eventlog of the PVS.

Customer gave me wrong LicenseServer Information or changed DNS.

I changed the license Server on PVS to the correct DNS and now it is working.

Please try also Telnet to license Server on Port 27000



Andrew Edwards Members
  • #17

Andrew Edwards
  • 33 posts

Posted 29 January 2015 - 10:18 AM

Brilliant, thanks for letting me know Alex. I haven't got any licensing setup which is why I had a sneaky suspicion. Thanks for posting back! 



Joerg Sonntag Members
  • #18

Joerg Sonntag
  • 4 posts

Posted 02 December 2015 - 08:32 AM

Thank you for this solution. I have the same problem at the moment. Could you please give me a hint, where/who I can change show/change the license server information on a existing PVS?

 

If I run the "Provisioning Services Configuration Wizard" again (with "farm is already configured" or "join existing farm"), I  don't get the "licence server" settings.

 

Thanks.



Frank Widak Members
  • #19

Frank Widak
  • 4 posts

Posted 20 March 2017 - 06:25 AM

I run in the same error - PVS 7.6 on 2012 R2 Server.

Target device boot works from both server, but failover and rebalance is not working.

System configuration is fine, I've checked everything.

 

Now the solution B) or how I found it:

I build the same cluster to reproduce and solve the problem by updating PVS. But upgrade was not needed.

When I imported my disks from productive cluster into my new one, I was able to failover and rebalance ... :blink:

 

So I'd the idea to do the same procedure on the productive server:

 

1.) Export vDisk

2.) Rename vDisk

3.) Import vDisk

4.) Add vDisk to target device

 

That solved the problem for me, the devices fail over and rebalance, hope that helps somebody.

 

BIT ROT :D