rguhr's Posts

We had a network outage yesterday because Storage vMotion was using too much bandwidth (to NFS Datastores).   Is there any way to limit Storage vMotion? For example, there is the possibility to set limits under System traffic. But there it says only vMotion is Storage vMotion included? (We don't want to limit "normal" NFS traffic)    

Hi, we setup multisite with two VCD instances. Customers now can view and manage their virtual data centers across sites with single login.   However, we have not found a way for customers to migrate workload from one site to another. The normal way within the same site would be the Move action on VMs (you need a target vAPP for this) but in the Move dialog you can only select vApps from the same site.   Does migration work differently for multisite or is there really no option for customers at the moment (without help from the service provider)? (The underlying vCenters can migrate workloads via Cross vCenter Migration)

Hi, we have two datacenter locations that use different L3 networks. We would like to be able to migrate VMs from one DC to the other. Networks: sideA-mgmt     (e.g. 10.10.1.0/24) sideA-vmotion (e.g. 10.10.2.0/24) sideB-mgmt     (e.g. 10.20.1.0/24) sideB-vmotion (e.g. 10.20.2.0/24) My first idea would be to use one combined vmkernel adapter for vmotion (hot) and provisioning (cold migration) traffic and set a route on both datacenter sides. I tested this solution and it worked but then I found this ~7 year old KB article that suggests routes for vmotion (and probably provisioning too?) are not official supported on default stack: https://kb.vmware.com/s/article/2108823 Is this still the case and I have to use own stacks for routeablity of vMotion/prov? My backup idea would be to create two vmkernel adapter one for vmotion and one for provisioning with their own stacks and then just put a default route into these own stacks / routing tables. (I would then need two IP addresses from the respective vMotion network, one for vmotion and one for prov. The gateway remains the same. Have not yet checked whether this makes problems with the routing table assignment. In the worst case I would need a third network per location.) Are both ways supported or only the own stack variant? Are routes ESXi (minor/major) upgrade safe? Regards, Robért

We use a shared catalog to distribute some basic ISOs to our customers. Unfortunately a lot of users don't remove the ISOs after they are done with it. To keep the shared catalog clean, we have to remove/update the ISOs from time to time. We have a powercli command to find VMs with a media connected on vSphere level:   Get-VM | Where-Object {$_.PowerState –eq "PoweredOn"}| Get-CDDrive | Where {$_.ISOPath -ne $null} | select Parent,IsoPath | format-table -autosize   But we need a version to this on VMware Cloud Director Level. Sometimes the VCD database thinks the media is still connected somewhere, but on vSphere level it's already gone. Get-Media and Get-CIVM apparently have no property that tracks that. It doesn't need to be a powercli cmdlet - a select query for the postgres database would also work