slim coral
Ok, I'm wondering if anyone has tried this. When replacing 2 5760 with 1 6060
Which is faster:
300TB on a 5660 takes 5 days to clone, 7 weeks to decommission. I assume the 5760 with 400TB will take 9-10 weeks. The 5760 would have ca. 400TB, so a cloning would take 7 or 8 days, perhaps. How long would a 6060 with ca. 400TB take to decommission?
rapid forge
I don't think you can clone 2 systems to one target
as you mentioned clone 5760 to 6060, then clone the 2nd 5760 to the same 6060
that is not possible
cloning process works one to one
if you want to use cloning two systems, you need to have twp target systems irrespective of the storahe space available in target systems
in your case, i think since you are trying to repalce 2 systems with one
things to consider are
ILM policy, make sure ILM policies are in line with number of storage nodes etc..
what is total number of nodes in the current grid?
what is teh current ILM and erase coding, 3+2, 2+1, etc...?
once we know the this information, i an provide next steps
time taken to expand a grid by adding storage node and removing existing nodes from the grid etc.. really depends on many factors. least time consuming can be, adding additional storage node/s, decommissioning one node at a time assuming there is enough usable space available in remaining storage nodes to accommodate distribution of data fro the node that is being decommissioned
assuming eraser coding can accommodate one less storage node in your case
slim coral
I don't think you really read my post correctly... the first 5760 is cloned to a 6060, this 6060 is then decommissioned and wiped clean, then reused to clone another 5760
rapid forge
thanks for clarifying. Let me research this but i don't have an answer to your question at present.
slim coral
ok
rapid forge
just out of curiosity, why do you want to clone 5760 to 6060, decommission 6060, then reuse to clone second 5760. What happens to the first 5760
just trying to understand the use case to see if there is an alternative to help accelerate the process to your end goal
slim coral
because i'm replacing the 5760's with just a single 6060 ... and decommissioning is incredibly long and often painful
rapid forge
so why clone first 5760 to 6060 and then decommission 6060 then wipe clean 6060? unless i am incorrectly reading your last post from yesterday
also how many storage nodes in your current grid of 5760 before you add 6060?
slim coral
because it may faster doing this way
i already have a number of 6060's... that's not really the point...
rapid forge
because it may faster doing this way
So you have a Grid with mix of 5760's and 6060's. You want to decommission the 5760's. In order to speed up the decommissioning process of the 5760's you want to clone 5760 to a new 6060 in the grid, then decommission 6060 wipe clean and follow the same process to decommission the second 5760. Yeah.?
rapid forge
because it may faster doing this way
I am probably limited with my knowledge of your environment and the need for deleting data from just about few storage nodes in the grid. I haven't faced this challenge in my current hosting scenarios therefor i recommend that before deleting/wiping data from the Grid please engage Systems Engineer from NetApp or professional services @ NetApp to review the scenario to full proof your plan of action.