Unsolved
This post is more than 5 years old
3 Posts
0
2899
WSM across VLANs
WSM across VLANs
Our test environment was a WSM server with Z00D win7 clients in the same VLAN. Everything was working fine with ~6 clients at once.
In production we moved the server (a vm) to a different vlan, and the clients are on a different vlan as well in a new building.
The clients are booting but after the first two we get all sorts of errors - mainly clients freezing at various stages during the boot/login process. The first client always boots normally/cleanly.
Anything we are missing?
mhaase-2015
200 Posts
0
June 4th, 2015 10:00
Thanks for sharing your findings and following the guidelines to find the root cause!
![](https://ip1.i.lithium.com/a5b20accd3cc9777705fbb4ad4bc20eb069f312e/68747470733a2f2f636f6d6d756e6974792e777973652e636f6d2f666f72756d2f696d616765732f736d696c6965732f736d696c652e676966)
So ifI understand correct, the problem occours only in a VM environment and "10G single mode fiber > Nexus 5596 (building2)" is part of the connection path.
Would be interesting to find out what exactly is causing the retries and if there is anything you can do to eliminate it.
At least you have a working reference environment with the physical WSM Server
dddds
3 Posts
0
June 4th, 2015 10:00
- Production server is same as in test
- Vmware ESXi 5.0.1 - using one VMXNET3 NIC
- no measurable lantency/bandwidth issues. Using Nexus 10G switches, with one hop to 10G 2960S w/ 1GB access ports.
- there is a small issue with the image here. normal size is ~300mb (win7) but sometimes ~900-1100mb.
- retries are astronomical. 30 to hundreds.
Two findings:
- Created a brand new physical WSM server connected to the same switch as the ESXi environment. Everything works fine.
- This second finding requires a bit more explanation. Path is ESXi cluster > Nexus 5596 (building1) > 10G single mode fiber > Nexus 5596 (building2) > Catalyst 2960S > Wyse Z00D. When changing the path to ESXi cluster > Nexus 5596 (building1) > catalyst 2960 1GB (building1) > Wyse Z00D, everything works fine.
So its not the network, because the physical server works fine. Its not vmware, because in the same building that works fine. Its VMware PLUS the network.
mhaase-2015
200 Posts
0
June 4th, 2015 10:00
- Is the production server the same as the one you used in your test environment or was it virtualized after the tests have been done?
- What virtualization vendor / version are you using for the WSM Server? We have got mixed results when virtualizing on VMware ESX5 or HyperV on hosts with multiple NICs. Use a physical server as reference to eliminate the virtualization layer causing hickups.
- Have you inspected the obvios performance counters on the server and measured the latency on the LAN and the available bandwith?
- How much data is transfered when a client is booting? What is the boot time of a single client? Inspect the client log files on the WSM server.
- What is the retry count on a booted client? It should be at 0, higher number indicates client must wait until the server is able to deliver the requested block (serverside ressource issue or slow network)
If you want to rise a support ticket at www.wyse.com/support, please complete and attach this form: ftp://support-emea.wyse.com/public/W...orm_111711.doc