If you're using the Virtual Storage Console, you can do this somewhat automatically with a reboot to finish the config change. What's your NetApp hardware config with respect to disks, aggregates, and controllers? Are there other workloads on these controllers or ESXi hosts?įor VMware on NetApp, there are a few NFS settings that NetApp recommends for your ESXi hosts. It protects you from IO storm but otherwise limits the performance while installation of a software or launching applications and so on.ĭo you limit IO per VM in your VDI environment to protect from rogue VM/s of this kind of event bringing the environment to crawl? I thought of SIOC but all the VMs are desktops and we cannot really differentiate high priority VMs and low priority VMs. We have played around and felt one way to tackle this would be limit IOs per VM, but it is two edged sword. However we would like to control the IOPS per VM and understand that there is a limit for IOPs under VM configuration of resources. I am trying to fix the source of the problem and that's where I want to focus my energy. I would like my fellow admins to chime in and say if they have witnessed something like this in their environment. It looks more like some kind of bug within browsers and flash plugin. We use systrack to analyze OS level activity and are not seeing a mass update of plugin. The situation stabilizes once the storm clears up. Essentially bringing the environment to a crawl for 5 - 10 minutes. Periodically we see some browsers (Chrome and Firefox) in combination with Flash and Flash Player Container generating 1000s of IOPS. We are in a View 5.0.1 in vSphere 5.0 environment running Windows 7 64bit.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |