Unsolved
1 Rookie
•
2 Posts
0
447
February 13th, 2023 15:00
Backup data sizing
Hello, We are planning to utilize cloud for LTR/archiving and need to some help with capacity planning. I was hoping to get some clarification on sizing reported by NMC but it is confusing. For the month of Jan, NMC is reporting below numbers:
Configuration: Single DD disk configured as AFTD in NW
Policies: Backup to dd (aftd) disk, and clone to LTO tapes
Data set: Mostly VMs, NAS, OS, and a few application (Incr/Full)
NMC Monthly Summary Report:
Amount of data: 1.7PB (Total size of backup on clients)
Target Size: 5TB (This is the actual data written on DD after de-dup)
Clone Size: 84TB
How is the clone size calculated? Shouldn't the clone size be equal or similar to the target size (actual data written on DD)? If that is the case, why is clone size reported in NMC way more than the target size? Whoever has implemented archiving on cloud, could you please share your experience and how did you do capacity planning?
Thank you in advance
barry_beckers
393 Posts
0
February 22nd, 2023 06:00
Using a DD as AFTD with Networker? Why?
Then all the benefits of using ddboost and client side deduplication are out of the window causing data always to be send over the network full blown instead of in deduped form? You would only be using target side dedupe on the DD itself and not client-side dedupe. So way more load on the network.
So don't you have or want a ddboost license on the DD? Works fantastic.
What version of NW are we dealing with here? And what "cloud" solution do you even want to go towards or use?
Cloned size would always state the amount of protected data, but nothing about how much (disk or tape) space is occupied to achieve that.
As generic rule of thumb we tend to say that 1TB of front-end protected data (so the largest full backup in last 60 days, what NW can report about for each NW client and VM using nsrcapinfo command) needs 1TB of disk-space on a DD when having at maximum of 1 month or retention. But that is with dedupe in play. Depending on what cloud solution you want to go towards, you might require more storage.
As total data (1.7PB) is way larger than the cloned data (84TB), this suggests not all data is cloned from the DD to tape? So all-in-all we are talking about needing to clone at max 84TB based on monthly numbers? However I'd actually make a calculation of ALL data in the clone pool(s), not just what was done last month as you are talking about LTR. You need to calculate everything what you would want to have in the cloud, which now still might be on tape. Even if not yet cloning into the cloud the current LTR data on tape, still in the near future you would have all new LTR data in the cloud and possibly need to arrange for the capacity for that?
If it indeed is mostly VM's, as they always report a backup as a full in NW, protected amount of data seems also way more than doing the same thing for the same client when using a NW agent installed on them and using a weekly full and the rest incr. backups. Hence dedupe achieved for VM backups seems better than with regular in-guest backups, but in fact the difference might not be that much different... Never did an exact 1-on-1 comparison really.