Zfs over iscsi plugin

The main problem is that multipathing with a lot of luns (1vm disk = 1 lun), is really painfull to manage. (remove, resize volumes...)

In past, I have also see multipath daemon blocking at 100% cpu when a lof of luns are present.

Best way is to use bonding for now.
 
spirit said:
The main problem is that multipathing with a lot of luns (1vm disk = 1 lun), is really painfull to manage. (remove, resize volumes...)

In past, I have also see multipath daemon blocking at 100% cpu when a lof of luns are present.

Best way is to use bonding for now.
Click to expand...

ok I see how this is working now, I’m looking at it from a different perspective of the hypervisor managing the multi-pathing and only having a few LUNs I.e. max 9tb in size with 30-50 VM per lun.

this would make the vm disk just a sparse file sitting on block storage formatted with something like lvm/ ext4 etc.

which is how VMware handles their iSCSI connections to a San or nas presenting block storage, I understand better now how KVM is or should I say qemu is managing this as each vm disk is a lun this is very similar to VMware‘s vvol technology https://kb.vmware.com/s/article/2113013

Ok now I understand this a lot better.

Do you know if normal iSCSI connectivity uses multi-path?

””Cheers
G
 
You must log in or register to reply here.
Top Bottom