We have 6 datastores in appvolumes -
3 for our standard VM pools
3 for our CAD VM pools
They all replicate to each other as some users move between CAD and STD pools.
When i update and appstack it takes forever for it to replicate to the other datastores. When i complete the appstack it takes even longer! Our CAD appstack is a 150GB template which is almost full now and that can take upto 4 hours!! A normal 20GB appstack can take around 1-2 hours.
We are running on 3PAR storage on SSD so this should be lightening quick.
I dont know if the issue is Appvolumes, Vcenter, ESXI, or even the 3PAR. It was all setup by another company before i had any involvement in it and i have found alot of things wrong with it over the past year.
If anyone has any bright ideas, i'd be very grateful because it now takes me about a day to update one appstack. It seems to be getting worse and worse as well.
AlexLSx7 For reference, one of our AppStacks that we update on a weekly/monthly basis is ~5.5GB, it takes less than 60 seconds to replicate when its the same physical site. When we replicate it to another site (10gbps link between the two sites), it does take a ~2 minutes. We run Nutanix on Dell hardware with ESXi on top. All storage is local to the hosts and is hybrid SSD/HDD on the nodes.
To take AppVolumes out of the mix, you could just copy some .vmdk's around yourself and see if its still slow. Technically AppVolumes is basically just doing the same thing anyway with the account you give permissions to vCenter for it.