thanks!
Any more cons?
thanks!
Any more cons?
Hi all,
I have 3 connection brokers and it seems that only one of them is randomly generating an issue. It happens that if a user tries to connect to a turned off desktop, the desktop doesn't turn on and the user receives an error regarding "desktop unavailable"
.
In the windows event viewer I can see an error id 104 stating "broker_machine_assigned_unavailable", in the debug log I can find the following error:
Result: error, Error Code: DESKTOP_LAUNCH_ERROR, Error Message: failed launching desktop: com.vmware.vdi.sessionclientapi.NoServersAvailableException$AssignedUnavailableException
If I manually turn on the desktop from the VC, the user connects just fine.
It happens randomly and it seems (but I'm not so sure) that it happens only with full desktops, the linked clones seem to be working fine.
Any hints?
Thanks everyone
It's always a good idea to open a support case when dealing with designs so that you know your environment will be supported. Its early but if I read that correctly your going to have two separate View environments, with separate everything component related, and sharing one single SQL Server and Host/Cluster?
Do you see an event in vCenter that shows where View attempts to power on the desktop? If so are there any errors on that event that would indicate why the desktop isn't powering on?
Hello mittim, thanks for your prompt answer.
on VC, I cannot see any entry in event viewer or in the events seen from the vsphere client. Moreover, if I look at the events related to the desktop I'm analyzing, I can see only the events related to a successful turn on (included the one I performed manually), but no events regarding any error.
Looks like the connection server doesn't contact VC at all...
is there any other place I can look into? logs somewhere?
This KB, VMware KB: Collecting diagnostic information for VMware View , covers how to create the log files. If you have narrowed it down to a specific CB problem and that CB is part of a replica group then it may be easier to just rebuild it.
I recently noticed that group policies that are assigned to my organization's user OU aren't applying. When I do a gpresult, it doesn't even indicate that the VM is looking at that OU at all. Are view machines supposed to load user policies like any other domain workstation?
Yes, there should not be any difference at all. Have a look at the OU and look for any exclusions from the GPO:s.
I found the cause, a silly mistake on my part. There's a computer policy that had loopback enabled but set to replace.
VM support suggested we restart all servers at the same time; however we couldn't schedule a time for this.
Our view servers all have IPV4 static IP Addresses assigned. Our DNS had entries for both IPV4 and IPV6 for each View server. The IPV6 address was being assigned via DHCP. The entry for IPV6 did not match between the View server and our DNS.
We removed the 'stale' entries from DNS and unchecked IPV6 from the NIC on each view server.
It now takes about 10 minutes for the admin page to become available.
What is a 'normal' amount of time for the admin page to be available? 10 minutes still seems long, but much better than the 30-45 minutes it was taking before.
Bandwidth is not an issue, this is for internal gigabit network. I'm hoping there is someone out there that has this working in their environment and can confirm setup / model numbers / anything else that can be useful for us.
Thanks
We are a CAD shop and we have tested vSGA and vDGA graphics acceleration with Horizon View 5.3. We are running Cisco UCS C240-M3 ESXi 5.1 U1 hosts with Nvidia K2 grid GPU cards. We are using Wyse P25 thin clients and love the performance we get with the hardware PCoIP provided by the Teradici chip and the vDGA GPU pass-thru capability now available with ESXi and View 5.3, but the fixed 1:1 relationship of VMs to GPUs is just not cost effective. We tried the vSGA shared model, but our most used apps used DirectX 11 and OpenGL 4.3, which are not supported in vSGA. The vSGA mode is really only useful for improving "normal" desktop users performance in a virtual desktop environment, for things like Office and web browsing and such, and not for any current generation CAD applications.
Any plans of a hybrid solution between these two modes? We really need the performance & compatibility that vDGA provides in the shared model that vSGA provides. Is this coming, and if so, how soon? We understand Citrix is already doing something like this and are considering switching platforms, but we really don't want to unless we have to.
- Chris
VMware's policy is not to comment on future features/products so you will probably not get any direct answers here.
You should reach out to your local VMware team to get a roadmap update from one of the SE:s.
Its correct that Citrix can do this with the Nvidia VGX technology (its more Nvidia then Citrix to be honest...), but then only XenServer is supported and you can then carve up the GPU:s in smaller chunks and use Nvidia driver with DX11 support.
My doubts about that VGX is that the GPU:s will maybe not be able to satisfy the users since they have to share them and you can over-allocate the resources.
(A K2 is basically 2 K5000 on a single PCI:e card so depending what your users have today it might not work for them to share one K5000.)
Would a few K1:s be a possible solution for you? Then you could double the amount of users per box compared to K2.
Another option is to use lower performance GPU:s in vDGA like the Q1000M/Q3000M or the older fermi-cards like Q4000.
// Linjo
Thanks Linjo! We've played around with K1 grid and Q4000 GPU cards and we're still looking for the sweet spot. We definitely like the performance of the K2 cards the best so far and not all of our shared users would need the performance concurrently, so we were hoping to get the best bang for our buck and still provide high power for individual users when needed. Our quest will continue and in the mean time, I'll hit up our VMware SE as you suggested. Thanks!
- Chris
hi there,
so whilst you might have a good network with loads of bandwidth, this doesn't mean that your USB redirection can use all of it. Unfortunately due to the way the USB standard works, it is synchronous - meaning that the computer sends a message to the device, the device replies, and then the computer says thanks. Then the next communication happens. Now, if this is a short wire then it can do this incredibly quickly - however for USB redirection the USB packets have to be sent over the network and whatever latency exists will significantly reduce the thruput you can achieve. More-so - because the packet sizes of USB are so small, you will find that it is sending small packets infrequently thus compounding the problem further.
So, I do believe that the bandwidth will be an issue. I also suspect (but haven't tried) that remoting these sorts of devices won't work either. Do try it, and post back your findings, but I don't expect it will work.
If it does work then before you roll out across many folks, then do please try varying the network conditions and simulate a real world network eg with 5ms latency and consider if any users will be connecting from home across a wifi and wan link and simulate that accordingly. as I believe this would definately lead to issues with the solution.
cheers
peterB
Hey there,
We are using 5.2 and Symantec 12.1 and have a number of linked clone desktops. We have AV definition updates 3 times a day and they linked clones' checkpoint files are growing rapidly and I believe its in relation to the AV definitions that they are getting. They seem to be renaming the old definitions afterwards and keeping 3 copies on the local linked clone disk, which is causing the checkpoint file to grow substantially as its differing from the Master each time by 300MB an update.
We are not in a position to refresh on logoff right now but am wondering if this is expected behaviour (???) as we find ourselves updating the Master on a a to-regurlar basis (because of the AV Def updates) and recomposing the Pools to get back the space on the Storage.
Appreciate any comments you have.
Cheers,
Paul.
Hi Sindhu!
Thanks for your answer - the Parent VM have to be placed in the according cluster, and it works!
So have a nice day, and
Best Regards,
Andre
Unfortunately it was not. I believe a bug was submitted for this issue.
By the way, which application are you having problems with?
Hi folks.
We have many international users. When connecting to VDI, which has been built with a UK locale, they may be using a French keyboard (this could be from any device of course, but I will focus on connecting from a Windows 7 laptop).
The users have the problem where the keyboard seems to revert to UK layout, but they are not familar with the UK layout and therefore where some keys are!
Two questions:
a) Should it not pass through the keyboard input method?
b) If not, what is the best way to overcome this - I know they can change the lanuage input on the VM they are connecting to... but is that the only/best workaround?
I expect this is a known issue, although I was having trouble finding info on my searches.
Can anyone confirm this behaviour?
Thanks.
Hi, thanks for the reply. Yeah we have done alot of scouting on the internet for this and have tried the Split Device and then not splitting the device. We can get this working fine under Windows 7 OS Client, we can get the device to be recongnised in Windows 8.1 it just doesnt passthough from the Windows 8.1 Client through to the Windows 7 VDI Machine.