The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

RAID info please -- Safe33 and Safe50 ???

Jack

Sr. Administrator
Staff member
Can somebody help me understand safe33 and safe50? Sounds like a virtual RAID1+0 but I cannot seem to follow the entire logic of how the info gets stored across each partition.
 

Dale Allyn

New member
Jack,

Did you come across this? It's something that I was reading as I was trying to figure this out too:

---
SAFE 33:

The SAFE33 storage policy creates two virtual volumes; one SAFE volume and one BIG volume, and should be used when you need the high reliability for some of your data (with the added overhead of mirroring) but you don’t need high reliability for the remainder of your data. SAFE33 reduces the cost of additional hard drives in operations where non-critical data could be lost without severe consequences. SAFE33 uses a SAFE volume that is mirrored across two hard drives to protect your critical data in the event a hard drive failure. If one drive fails the SAFE volume is retrievable although the BIG volume is not. When you replace the failed drive, the SAFE volume is automatically rebuilt on to the replacement drive. For example, if you are using a video editing application that stores the primary source data and also uses some temporary storage for editing, you need protected storage that is offered by SAFE for the primary source data, but you do not need protected storage for the temporary data. Therefore, the combination of SAFE and BIG would be the most efficient utilization of your available storage capacity. If either hard drive fails the primary data stored on the SAFE volume would still be available whereas the temporary data stored on the BIG volume would be lost. The size of the SAFE volume of a SAFE33 policy will be one-third of the size of one hard drive (if they are equal) or one-third of the size of the smaller (if they are not equal.) The size of the BIG volume will be the combination of all remaining capacities.

Example: In Figure below, assume that Drives A and B are 300 GB each. When the SAFE33 Storage Policy is selected, the resulting virtual volumes will include SAFE volume of 100 GB (1/3 of 300 GB) and a BIG volume of 400 GB (the remaining capacity after allocating 100 GB from each of the hard drives).

SAFE33 storage policy sample configuration

SAFE 50:

The SAFE50 storage policy creates two virtual volumes; one SAFE volume and one BIG volume, and should be used when you need the high reliability for some of your data (with the added overhead of mirroring) but you don’t need high reliability for the remainder of your data. SAFE50 reduces the cost of additional hard drives in operations where non-critical data could be lost without severe consequences. SAFE50 uses a SAFE volume that is mirrored across two hard drives to protect your critical data in the event a hard drive failure. If one drive fails the SAFE volume is retrievable although the BIG volume is not. When you replace the failed drive, the SAFE volume is automatically rebuilt on to the replacement drive. For example, if you are using a video editing application that stores the primary source data and also uses some temporary storage for editing, you need protected storage that is offered by SAFE for the primary source data, but you do not need protected storage for the temporary data. Therefore, the combination of SAFE and BIG would be the most efficient utilization of your

available storage capacity. If either hard drive fails the primary data stored on the SAFE volume would still be available whereas the temporary data stored on the BIG volume would be lost. The size of the SAFE volume of a SAFE50 policy will be one-half of the size of one hard drive (if they are equal) or one-half of the size of the smaller (if they are not equal). The size of the BIG volume will be the combination of all remaining capacities.

Example: In Figure below, assume that Drives A and B are 300 GB each. When the SAFE50 Storage Policy is selected, the resulting virtual volumes will include SAFE volume of 150 GB (1/2 of 300 GB) and a BIG volume of 300 GB (the remaining capacity after allocating 150 GB from each of the hard drives).

SAFE50 storage policy sample configuration

Note:

When using a SATA host connection, you must have a PM (Port Multiplier) aware host adapter when using SAFE33 (or SAFE50) on the top level node of a cascaded configuration so that ALL volumes can be detected by the host. If your SATA host adaptor is not PM aware, then ONLY the SAFE volume will be detected and the BIG volume will not be accessible. No such limitation exists when using a USB host connection. For subordinate nodes in a cascaded configuration, it is possible to configure a SAFE33 (or SAFE50)storage policy, although you will only see the SAFE volume from that node. Therefore, the SAFE33 (or SAFE50) storage policy should only be used at the top-level node of a cascaded configuration. In this mode, the Schedule/Verify option is enabled.
---
It was on the cooldrives site: http://www.cooldrives.com/esandusb20du.html
 

Jack

Sr. Administrator
Staff member
that's the standard explanation on all the sites. What I don't understand is the math behind them as the two volumes do not match up in total volume; how can 1/3 of a drive(s) provide safe mirror the other 2/3?
 

Dale Allyn

New member
As I'm understanding it, and maybe this is obvious, but in SAFE33 we have 1/3 of our data which is safe by mirrored copy with the remaining data not protected. Yes, there is an imbalance to the math. IOW 2/3 of your data is not safe, but it requires less storage space and only protects the most critical data.

For me, if I understand it correctly, it's of no use, but is of interest to those who don't need to protect versioning, for example.

I'm still absorbing this too, but had decided that it was not good for my photo stuff because I want complete data mirroring.
 

Dale Allyn

New member
This was about the only thing I found in my heap of reading that sort of clarified the math:

SAFE33 and SAFE50 are hybrid modes that deliver capacity plus mirroring with two drives, but they don't employ striping like RAID 0+1 does so there's no performance increase over, say, JBOD or RAID 1.

In the case of SAFE33, to give an example, you'll see 67% of the capacity of both discs combined as a single logical volume, and the rest as a second, mirrored volume. Thus, if you install a pair of 500GB drives, which would combine for 1000GB, SAFE33 offers 670GB for vanilla storage and uses the other 330GB for mirroring—you'd see 115 of the latter because the rest is redundant.
---
It just struck me as a sort of seamless partitioning, and I've not grasped an advantage for anything I do.
 
Last edited:

Jack

Sr. Administrator
Staff member
Thanks Dale, that clarified it for me. What I thought I was missing was some added value to it, like performance plus redundancy. You've arrived at and confirmed the same conclusion I did -- with no performance gain and limited redundancy, it isn't of much value :)

thanks for your help!
 

etrigan63

Active member
Jack, you may want to take a look at the Drobo. This device is RAID for the rest of us. Four trayless bays allow you to add storage on the fly. There is a network module that turns it into a NAS. Pretty fast from what I've seen. I am trying to get one in for review. You can also do this with a Windows Home Server like the MediaSmart Server from HP. I built one from scratch and it works pretty well too. The WHS solution has the advantage of mirroring itself automatically and (with a JungleDisk plugin) backing itself up offsite.
 
Top