Hi All,
Now, I study 3par
I found it very complex storage configuration, so I need some help from you....
I need know what is deference between RAW cap, Usable Gib and RAW Gib??
If I have 3 tiers NL + SSD + SAS, is data write first in SSD??
when data change place (from tier to another tier)??
what the deference between Optimization Software Suite v1 and v2?
what is the deference between (BE IOPS and FE IOPS) and (ME MBPS and FE MBPS)??
what is a chunklets, magazine and CPVV??
what is deference between HA DISK and HA Enclosure?
I need pdf for configure 3par.
Thanks & Best Regards
Ahmed
Tiering?? RAW cap vs usable GiB vs RAW GiB vs InformOS) ??
-
Eng.AhmedAmin
- Posts: 10
- Joined: Mon Apr 14, 2014 11:01 am
Re: Tiering?? RAW cap vs usable GiB vs RAW GiB vs InformOS)
Raw capacity GB = sum of drive manufacturer advertised capacity base 10, 1000MB per GB
Raw GiB = sum of actual raw capacity before raid overhead base 2, 1024MB per GB
Usable GiB = sum of actual usable capacity after raid overhead base 2, 1024MB per GB
3 tiers, new writes will go to whichever tier the volume was created in, typically the middle tier as this has the best blend of capacity and performance. If it went to SSD then you have the potential to fill SSD quickly, if it goes to NL then performance could be an issue.
Movement is based on a schedule, transient peaks are handled by cache.
Opt suite v1 = Dynamic optimization, Adaptive optimization & Peer motion
Opt suite v2 = All of the above plus Priority optimization (QOS)
BE = Back end, what the disks are doing including raid overheads etc (performance) MB/s or IOps
FE = Front end, what the hosts, applications are seeing (performance) MB/s or IOps
Chunklet = logical container (1GB) created from physical disk to house raid & data
LD = Logical disk a collection of chunklets per controller, backing virtual volumes and assembled based on CPG policy
Mag = 10000 series a sled of 4 disks, 7000 series a single disk.
CPG = Common provisioning group, policy or template used to direct volume properties and growth
CPVV = Common provisioned virtual volume, non thin provisioned volume drawn from a CPG
TPVV = Thin provisioned virtual volume drawn from a CPG
HA Cage - data is automatically striped vertically across disk enclosures providing the ability for volumes to survive the loss of an entire disk enclosure (24 disks 7000 series or 40 disks 10000 series)
HA Mag - no guarantee of the above i.e the same as everyone else's raid
Good luck, but I'd start with the 3PAR concepts guide http://lmgtfy.com/?q=hp+3par+concepts+guide
Step by step guide here (a bit old) http://lmgtfy.com/?q=3cvguy+3par+stores ... stallation
Raw GiB = sum of actual raw capacity before raid overhead base 2, 1024MB per GB
Usable GiB = sum of actual usable capacity after raid overhead base 2, 1024MB per GB
3 tiers, new writes will go to whichever tier the volume was created in, typically the middle tier as this has the best blend of capacity and performance. If it went to SSD then you have the potential to fill SSD quickly, if it goes to NL then performance could be an issue.
Movement is based on a schedule, transient peaks are handled by cache.
Opt suite v1 = Dynamic optimization, Adaptive optimization & Peer motion
Opt suite v2 = All of the above plus Priority optimization (QOS)
BE = Back end, what the disks are doing including raid overheads etc (performance) MB/s or IOps
FE = Front end, what the hosts, applications are seeing (performance) MB/s or IOps
Chunklet = logical container (1GB) created from physical disk to house raid & data
LD = Logical disk a collection of chunklets per controller, backing virtual volumes and assembled based on CPG policy
Mag = 10000 series a sled of 4 disks, 7000 series a single disk.
CPG = Common provisioning group, policy or template used to direct volume properties and growth
CPVV = Common provisioned virtual volume, non thin provisioned volume drawn from a CPG
TPVV = Thin provisioned virtual volume drawn from a CPG
HA Cage - data is automatically striped vertically across disk enclosures providing the ability for volumes to survive the loss of an entire disk enclosure (24 disks 7000 series or 40 disks 10000 series)
HA Mag - no guarantee of the above i.e the same as everyone else's raid
Good luck, but I'd start with the 3PAR concepts guide http://lmgtfy.com/?q=hp+3par+concepts+guide
Step by step guide here (a bit old) http://lmgtfy.com/?q=3cvguy+3par+stores ... stallation
Last edited by Cleanur on Wed Apr 23, 2014 5:11 pm, edited 3 times in total.
Re: Tiering?? RAW cap vs usable GiB vs RAW GiB vs InformOS)
I nominate Cleanur's post for sticky status =)