3PAR StoreServ Utility Storage - Lake Solutions AG
3PAR StoreServ Utility Storage - Lake Solutions AG
3PAR StoreServ Utility Storage - Lake Solutions AG
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
HP <strong>Storage</strong> Event<br />
<strong>3PAR</strong> <strong>StoreServ</strong> <strong>Utility</strong> <strong>Storage</strong><br />
Alfio Lazzari<br />
Geschäftsführer
Agenda<br />
14:30 – 14:45 Begrüssung Alfio Lazzari (LAKE)<br />
14:45 – 15:15 Machen Sie Ihre Daten schlank<br />
Holen Sie mehr aus Ihrer Diskkapazität mit der <strong>3PAR</strong> <strong>StoreServ</strong> 7000 Thin Technologie<br />
15:15 – 15:45 Mit HP Technologie bleiben Sie im Driving Seat<br />
Fredy Wismer (LAKE)<br />
Verkürzen Sie die RPO und RTO Zeiten durch den Einsatz von SnapShot & Datenreplizierung Patrick Pletscher (LAKE)<br />
15:45 – 16:00 Ihr persönliches <strong>StoreServ</strong> <strong>Storage</strong> System<br />
Wir ermitteln für Sie die optimale <strong>StoreServ</strong> Konfiguration durch gezielte Datenerhebung<br />
16:00 – 16:30 Daten schnell und sicher gespeichert mit<br />
HP StoreOnce Backup2Disk<br />
16:30 – 17:30 Pause / Apéro<br />
17:30 – 19:30 Kino Start – Iron Man 3<br />
Daniel Estermann (LAKE)<br />
René Loser (HP)
Machen Sie Ihre Daten schlank<br />
& fit<br />
Hr. Fredy Wismer<br />
Solution Architect
Agenda<br />
• Produkteübersicht<br />
• Funktionsübersicht<br />
• Hardware und Software – Lizenzierung<br />
• Fitness anhand eines Beispiels der LAKE Cloud
HP <strong>Storage</strong> Array Positioning<br />
<strong>Storage</strong><br />
Consolidation<br />
P2000 MSA<br />
StoreSure<br />
Virtual IT<br />
P4000 LeftHand<br />
StoreVirtual<br />
Application<br />
Consolidation<br />
P6000 EVA<br />
<strong>Utility</strong> <strong>Storage</strong><br />
P10000 <strong>3PAR</strong><br />
<strong>StoreServ</strong><br />
Mission Critical<br />
Consolidation<br />
P9500<br />
Architecture Dual Controller Scale-out Cluster Dual Controller Mesh-Active Cluster Fully Redundant<br />
Connectivity FC, SAS, iSCSI iSCSI, FC FC, iSCSI, FCoE FC, iSCSI, (FCoE planned) FC, FCoE<br />
Performance 30K random read IOPs ;<br />
1.5GB/s sequential reads<br />
35K random read IOPs<br />
2.6 GB/s sequential reads<br />
55K random read IOPS<br />
1.7 GB/s sequential reads<br />
> 390K random read IOPs;<br />
> 14 GB/s sequential reads<br />
>1000K random read IOPS<br />
> 10GB/s sequential reads<br />
Application<br />
Sweet spot<br />
SMB , Enterprise ROBO,<br />
consolidation/<br />
virtualization<br />
Server attach, Video<br />
surveillance<br />
SMB, ROBO and Enterprise<br />
– Virtualized inc VDI ,<br />
Microsoft apps, ITaaS<br />
BladeSystem SAN (P4800)<br />
Enterprise - Microsoft,<br />
Virtualized, OLTP<br />
Enterprise and Service<br />
Provider , ITaaS, Utilities,<br />
Cloud, Virtualized<br />
Environments, OLTP, Mixed<br />
Workloads<br />
Large Enterprise - Mission<br />
Critical w/Extreme<br />
availability, Virtualized<br />
Environments, Multi-Site<br />
DR<br />
Capacity<br />
600GB – 192TB;<br />
6TB average<br />
7TB – 768TB;<br />
72TB average<br />
2TB – 480TB;<br />
36TB average<br />
5TB – 1600TB;<br />
120TB average<br />
10TB – 2000 TB;<br />
150TB average<br />
Key features<br />
Price / performance<br />
Controller Choice<br />
Replication<br />
Server Attach<br />
All-inclusive SW<br />
Multi-Site DR included<br />
Virtualization<br />
VM Integration<br />
Virtual SAN Appliance<br />
Ease of use and Simplicity<br />
Integration/Compatibility<br />
Multi-Site Failover<br />
Multi-tenancy<br />
Efficient Thin Technologies<br />
Performance<br />
Autonomic Tiering and<br />
Management<br />
Constant Data Availability<br />
Heterogeneous<br />
Virtualization<br />
Multi-site Disaster<br />
Recovery<br />
Application QOS (APEX)<br />
Smart Tiers<br />
OS support<br />
Windows, vSphere, HP-UX,<br />
Linux, OVMS, Mac OS X,<br />
Solaris, Hyper-V<br />
vSphere. Windows, Linux,<br />
HP-UX, MacOS X, AIX,<br />
Solaris, XenServer<br />
Windows, VMware, HP-UX,<br />
Linux, OVMS, Mac OS X,<br />
Solaris, AIX<br />
vSphere, Windows, Linux,<br />
HP-UX, AIX, Solaris, OVMS<br />
All major OS’s including<br />
Mainframe and Nonstop
HP <strong>3PAR</strong> <strong>StoreServ</strong> <strong>Storage</strong> Familie<br />
F200<br />
Ein OS, Management Console und Set von Software Features<br />
F400<br />
<strong>StoreServ</strong><br />
7200<br />
<strong>StoreServ</strong><br />
7400<br />
<strong>StoreServ</strong><br />
10400<br />
<strong>StoreServ</strong><br />
10800<br />
Controller Nodes 2 2 – 4 2 2 - 4 2 – 4 2 – 8<br />
Fibre Channel Host Ports<br />
Optional 1Gb iSCSI Ports<br />
Optional 10Gb iSCSI Ports<br />
Optional 10Gb FCoE Ports<br />
Built-in IP Remote Copy<br />
Ports<br />
0 – 12<br />
0 – 8<br />
NA<br />
NA<br />
2<br />
0 – 24<br />
0 – 16<br />
NA<br />
NA<br />
2 – 4<br />
4 – 12<br />
NA<br />
0 – 4<br />
0 – 4<br />
2<br />
4 – 24<br />
NA<br />
0 – 8<br />
0 – 8<br />
2 – 4<br />
0 – 96<br />
NA<br />
0 – 16<br />
0 – 16<br />
2 – 4<br />
0 – 192<br />
NA<br />
0 – 32<br />
0 – 32<br />
2 – 4<br />
GBs Cache 20 20 – 40 24 32 – 64 96 – 192 192 – 768<br />
Disk Drives 16 – 192 16 – 384 8 – 144 8 – 480 16 - 960 16 - 1920<br />
Available<br />
Drive<br />
Types<br />
SLC SSD<br />
(max/array)<br />
15krpm<br />
10krpm<br />
7.2krpm<br />
100, 200GB<br />
(32)<br />
300, 600GB<br />
450, 900GB<br />
1, 2TB<br />
100, 200GB<br />
(64)<br />
300, 600GB<br />
450, 900GB<br />
1, 2TB<br />
100, 200GB<br />
(120)<br />
300GB<br />
450, 900GB<br />
2, 3TB<br />
100, 200GB<br />
(240)<br />
300GB<br />
450, 900GB<br />
2, 3TB<br />
100, 200GB<br />
(256)<br />
300, 600GB<br />
450, 900GB<br />
1, 2TB<br />
100, 200GB<br />
(512)<br />
300, 600GB<br />
450, 900GB<br />
1, 2TB<br />
Read throughput MB/s<br />
IOPS (true backend IOs)<br />
1,300<br />
34,400<br />
2,600<br />
76,800<br />
2,500<br />
150,000<br />
4,800<br />
320,000<br />
6,500<br />
180,000<br />
13,000<br />
360,000<br />
SPC-1 Benchmark IOPS NA 93,050 NA NA NA 450’213
Constant evolution<br />
May 99<br />
<strong>3PAR</strong> founded<br />
with 5 employees<br />
July 01<br />
<strong>3PAR</strong> secures $100 Million in<br />
third-round financing<br />
June 02<br />
<strong>3PAR</strong> <strong>Utility</strong> <strong>Storage</strong> and Thin<br />
Provisioning launch in the US<br />
and Japan<br />
November 07<br />
<strong>3PAR</strong> IPO<br />
Introduction of:<br />
Virtual Domains,<br />
native iSCSI support<br />
HP <strong>3PAR</strong> Geschichte<br />
March 10<br />
Introduction of:<br />
Adaptive Optimization,<br />
Recovery Manager for VMware<br />
September 10<br />
<strong>3PAR</strong> acquired by HP<br />
November 11<br />
New SPC-1 world<br />
record released:<br />
450’213 IOPS<br />
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012<br />
2013<br />
December 00<br />
Bring-up of the Gen 1<br />
<strong>3PAR</strong> ASIC<br />
September 02<br />
General Availability of<br />
the InServ S-Class<br />
<strong>Storage</strong> Server<br />
August 06<br />
Introduction of:<br />
E-Class midrange<br />
<strong>Storage</strong> Server<br />
May 05<br />
<strong>3PAR</strong> introduces Dynamic<br />
Optimization and Recovery<br />
Manager<br />
November 09<br />
InForm OS v2.3.1<br />
with many new<br />
features<br />
April 09<br />
Introduction of:<br />
F-Class the first quad<br />
controller midrange<br />
array<br />
September 08<br />
Introduction of:<br />
T-Class with Gen 3 ASIC -<br />
the first Thin Built in<br />
storage array<br />
August 11<br />
Introduction of:<br />
V-Class w. Gen 4 ASIC,<br />
InForm v3.1.1,<br />
Peer Motion<br />
December 12<br />
Introduction of:<br />
New <strong>StoreServ</strong> 7000,<br />
<strong>3PAR</strong> OS 3.1.2,<br />
Peer Persistence,<br />
Persistent Ports
HP <strong>3PAR</strong> Kunden reduzieren TCO um 50%<br />
HP <strong>3PAR</strong> führend bezüglich Effizienz<br />
Thin Technologies<br />
Tiering Optimized<br />
Green<br />
Tier 0<br />
Tier 1<br />
Tier 2<br />
Reduktion der benötigen<br />
Kapazität um mindestens<br />
50% - garantiert *<br />
* Mehr über HPs Thin Garantie unter http://www.hp.com/storage/getthin<br />
Tiering gleicht Kosten aus<br />
zwischen CHF/GB<br />
und CHF/IOP<br />
Einsparung bei Strom und<br />
Kühlung um mindestens<br />
50%<br />
wie auch bei Investitionsund<br />
Betriebskosten
HP <strong>3PAR</strong> Merkmale<br />
<strong>3PAR</strong> Thin Technologies<br />
Thin Provisioning, Thin Conversion, Stay Thin : sehr Schlank <br />
<strong>3PAR</strong> Optimization Technologies<br />
Adaptive und Dynamic Optimization Tiering<br />
<strong>3PAR</strong> Persistent Technologies<br />
Service Levels auch unter ausfall Situationen gewährleisten<br />
<strong>3PAR</strong> Federation Technologies<br />
Federated data mobility and availability<br />
<strong>3PAR</strong> Mesh-Active Architecture<br />
Multi-tenant, autonomic and efficient storage
HP <strong>3PAR</strong> ASIC<br />
Hardware Unterstützte Optimierung<br />
Thin Built in Zero Detect<br />
Fast RAID 10, 50 & 60<br />
Rapid RAID Rebuild<br />
Integrated XOR Engine<br />
Tightly-Coupled Cluster<br />
High Bandwidth, Low Latency<br />
Interconnect<br />
Mixed Workload<br />
Independent Metadata and Data<br />
Processing
HP <strong>3PAR</strong> Thin Technologies<br />
Start Thin Get Thin Stay Thin<br />
Traditional<br />
Provisioning<br />
<strong>3PAR</strong> Thin<br />
Provisioning<br />
4TB<br />
4TB<br />
1TB<br />
1TB<br />
Thin Provisioning<br />
• No pool management or<br />
reservations<br />
• No professional services<br />
• Fine capacity allocation units<br />
• Variable QoS for snapshots<br />
Thin Conversion<br />
• Eliminate time & complexity of getting thin<br />
• Open, heterogeneous migrations for any<br />
array to <strong>3PAR</strong><br />
• Service levels preserved during conversion<br />
Thin Persistence<br />
• Free stranded capacity<br />
• Automated reclamation based on<br />
T10 write same or unmap operations<br />
• Snapshots and Remote Copies stay thin<br />
Buy up to 75% less<br />
storage capacity<br />
Reduce Tech Refresh<br />
Costs by up to 60%<br />
Thin Deployments<br />
Stay Thin Over time
<strong>3PAR</strong> Mixed Workload Verarbeitung<br />
Unterschiedliche I/O Muster<br />
= control information (metadata)<br />
= data<br />
Heavy throughput<br />
workload applied<br />
I/O Processing : Traditionelles <strong>Storage</strong><br />
Heavy transaction<br />
workload applied<br />
Host<br />
interface<br />
Unified Processor<br />
and/or Memory<br />
Kleine IOs warten bis grosse IOs fertig verarbeitet sind<br />
Disk<br />
interface<br />
disk<br />
hosts<br />
I/O Verarbeitung bei HP <strong>3PAR</strong> Speichersystemen<br />
Heavy throughput<br />
workload sustained<br />
<strong>3PAR</strong> ASIC &<br />
Memory<br />
Heavy transaction<br />
workload sustained<br />
Host<br />
interface<br />
Control Processor &<br />
Memory<br />
Metadaten und Applikationsdaten werden getrennt verarbeitet<br />
Disk<br />
interface<br />
disk
<strong>3PAR</strong> <strong>StoreServ</strong> 7000 Architektur<br />
Kostengünstig, Skalierbar, Verfügbar, “Mesh Archtiektur” und fit<br />
<strong>StoreServ</strong> 7200<br />
mit 2 Knoten<br />
7400<br />
with 2 nodes<br />
<strong>StoreServ</strong> 7400<br />
mit 4 Knoten<br />
Write-Cache nicht<br />
beeinträchtigt<br />
Host Ports<br />
Persistent Write-Cache Spiegelung<br />
• Keinen Einfluss auf Schreib-Performance<br />
• Ab 4 oder mehr <strong>StoreServ</strong> Knoten<br />
7400<br />
Data Cache<br />
Disk Ports<br />
10400, 10800
Virtualisierung<br />
Traditional Array<br />
• Each RAID level requires dedicated drives<br />
• Dedicated spare disk required<br />
• Limited single LUN performance<br />
HP <strong>3PAR</strong><br />
• All RAID levels can reside on same drives<br />
• Distributed sparing, no dedicated spares<br />
• Built-in wide-striping based on Chunklets<br />
0 1 2 3 4 5 6 7<br />
R1 R1 R5 R6 R6 R1 R5 R5<br />
Traditional Controllers<br />
0 1 2 3 4 5 6 7<br />
R1 R1 R5 R6 R6 R1 R5 R5<br />
<strong>3PAR</strong> <strong>StoreServ</strong> Controllers<br />
RAID5 Set<br />
RAID1<br />
RAID6 Set<br />
LUN 6<br />
LUN 5<br />
LUN 7<br />
RAID1 Set<br />
LUN 4<br />
LUN 3<br />
RAID5 Set<br />
Spare<br />
Physical drives<br />
LUN 0<br />
LUN 1<br />
LUN 2<br />
Spare
Chunklets<br />
Chunklets<br />
HP <strong>3PAR</strong> Dynamic Optimization<br />
Systemauslastung bzw. Leistung optimieren durch “Rebalancing”<br />
Verteilung nach zwei Disk Ausbauten<br />
Verteilung nach Dynamic Optimization<br />
600<br />
Before Dynamic Optimization<br />
Free<br />
Used<br />
600<br />
After Dynamic Optimization<br />
Free<br />
Used<br />
500<br />
500<br />
400<br />
400<br />
300<br />
300<br />
200<br />
200<br />
100<br />
0<br />
1 20 39 58 77 96<br />
Physical Disks<br />
REBALANCE<br />
100<br />
0<br />
1 20 39 58 77 96<br />
Physical Disks
HP <strong>3PAR</strong> Dynamic und Adaptive Optimization<br />
Manuelles oder automatisches Tiering<br />
<strong>3PAR</strong> Dynamic<br />
Optimization<br />
<strong>3PAR</strong> Adaptive<br />
Optimization<br />
Tier 0<br />
CPG 1<br />
CPG A<br />
Tier 1<br />
CPG 2<br />
CPG B<br />
Tier 2<br />
CPG 3<br />
CPG C<br />
- Region (Datenblöcke)<br />
LUN Verschiebung<br />
zwischen Tiers<br />
Sub-LUN Block Verschiebung zwischnen<br />
Tiers anhand eines Regelwerks
HP <strong>3PAR</strong> Dynamic Optimization - Fallbeispiel<br />
Service Level möglichst günstig über ganzen Lebenszyklus gewährleisten<br />
10TB net<br />
10TB net<br />
10TB net<br />
10TB net<br />
(20TB<br />
effektiv)<br />
(13TB<br />
effektiv)<br />
(12TB<br />
effektiv)<br />
(13TB<br />
effektiv)<br />
RAID 10<br />
300GB 15K Disks<br />
RAID 50 (4+1)<br />
300GB 15K Disks<br />
RAID 50 (7+1)<br />
900GB 10K Disks<br />
RAID 60 (8+2)<br />
3TB 7.2K Disks<br />
50% Platzeffizenz<br />
75% Platzeffizenz<br />
88% Platzeffizenz<br />
80% Platzeffizenz<br />
CAPEX und OPEX Einsparung über gesammten Lebenszyklus der Daten
<strong>3PAR</strong> <strong>StoreServ</strong> 7000<br />
HP <strong>3PAR</strong> <strong>StoreServ</strong> 7200 HP <strong>3PAR</strong> <strong>StoreServ</strong> 7400<br />
Controller Nodes 2 2 4*<br />
Max. SFF Disks 144 240 480<br />
Cache 24 GB 32 GB 64GB<br />
8Gbit/s FC ports<br />
total (built-in/optional)<br />
12 (4/8) 12 (4/8) 24 (8/16)<br />
optional 10Gbit/s iSCSI (FCoE) 4 4 8<br />
Built-in IP remote copy port 2 2 4<br />
Controller Enclosures<br />
2U jeweils mit 24 SFF Disk Steckplätzen<br />
Service Processor (Alarmierung, HW Service)<br />
Drive Enclosures<br />
2U mit 24 SFF und/oder 4U mit 24 LFF Disks<br />
1 1 2<br />
Virtuell (VM) oder dedizierte Hardware (HW)<br />
0 to 5 0 to 9 0 to 18
<strong>3PAR</strong> <strong>StoreServ</strong> 7000 Kontroller<br />
Rückansicht<br />
Front Ansicht<br />
Node 1<br />
Node 0<br />
Eth Mgmt. Port<br />
2x built-in 8Gbit/s FC<br />
2x 4-lane 6Gbit/s SAS for drive chassis connection<br />
Optional PCIe card<br />
- 4x 8GB/s HBA or<br />
- 2x 10Gbit/s CNA<br />
Controller interconnect (7400 only)<br />
Built-in Eth Remote Copy Port
4-node System mit Interconnect Verkabelung<br />
<strong>3PAR</strong> <strong>StoreServ</strong> 7400 4-node System<br />
Node 3<br />
Node 2<br />
Node 1<br />
Node 0
<strong>3PAR</strong> 7000 Disk Enclosure<br />
2½ und 3½ Festplatten Enclosures mit je 24 Slots – frei kombinierbar<br />
2 Höheneinheiten (HE)<br />
mit 24 Small Form Factor (SFF) Steckplätzen<br />
4HE mit 24 LFF Disk Steckplätzen
HP <strong>3PAR</strong> 7000 Lizenzierung<br />
Operating System Suite<br />
Optimization Suite<br />
Replication Suite<br />
Security Suite<br />
Reporting Suite<br />
Thin Provisioning<br />
Thin Copy Reclamation<br />
Thin Persistence<br />
Thin Conversion<br />
Full Copy<br />
<strong>3PAR</strong> OS Administration Tools<br />
CLI client<br />
SNMP<br />
Virtual SP<br />
Online Import license (180 days)<br />
SmartStart<br />
System Tuner<br />
Host Explorer<br />
Multi Path IO SW<br />
VSS Provider<br />
Rapid Provisioning<br />
Autonomic Groups<br />
Autonomic Replication Groups<br />
Autonomic Rebalance<br />
Scheduler<br />
LDAP Support<br />
Access Guard<br />
Host Personas<br />
Persistent Cache<br />
Persistent Ports<br />
Management Console<br />
Web Services API<br />
SMI-S<br />
Dynamic Optimization<br />
Adaptive Optimization<br />
Peer Motion<br />
Application Suite for<br />
VMware<br />
Recovery Manager for<br />
vSphere<br />
VASA<br />
vCenter plug-in<br />
vSphere Host Explorer<br />
Virtual Copy (VC)<br />
Remote Copy (RC)<br />
Peer Persistence<br />
Application Suite for<br />
MS Exchange<br />
Recovery Manager<br />
for MS Exchange<br />
Virtual Domains<br />
Virtual Lock<br />
Application Suite for<br />
MS SQL<br />
Recovery Manager<br />
for MS SQL<br />
System Reporter<br />
<strong>3PAR</strong>Info<br />
Application Suite for<br />
Oracle<br />
Recovery Manager<br />
for Oracle<br />
Basis<br />
Optional<br />
Optional<br />
– Lizenziert pro Disk – Lizenz „offen“ für 7200 bei 48, 7400 bei 168 Disks<br />
– Lizenziert pro Disk – Lizenz „offen“ für 7200 bei 48, 7400 bei 168 Disks<br />
– Lizenziert pro <strong>3PAR</strong> System
Fragen & Antworten
Und weshalb fit?<br />
Fallbeispiel: Auflösen<br />
eines VMware Snapshots<br />
• SATA Disks (NL) Pool (CPG)<br />
• Über mehrere Stunden<br />
>200 IOPs pro Disk (PD)<br />
• Alarmierung automatisch<br />
• HP Support<br />
• Kunde
Agenda<br />
14:30 – 14:45 Begrüssung Alfio Lazzari (LAKE)<br />
14:45 – 15:15 Machen Sie Ihre Daten schlank<br />
Holen Sie mehr aus Ihrer Diskkapazität mit der <strong>3PAR</strong> <strong>StoreServ</strong> 7000 Thin Technologie<br />
15:15 – 15:45 Mit HP Technologie bleiben Sie im Driving Seat<br />
Fredy Wismer (LAKE)<br />
Verkürzen Sie die RPO und RTO Zeiten durch den Einsatz von SnapShot & Datenreplizierung Patrick Pletscher (LAKE)<br />
15:45 – 16:00 Ihr persönliches <strong>StoreServ</strong> <strong>Storage</strong> System<br />
Wir ermitteln für Sie die optimale <strong>StoreServ</strong> Konfiguration durch gezielte Datenerhebung<br />
16:00 – 16:30 Daten schnell und sicher gespeichert mit<br />
HP StoreOnce Backup2Disk<br />
16:30 – 17:30 Pause / Apéro<br />
17:30 – 19:30 Kino Start – Iron Man 3<br />
Daniel Estermann (LAKE)<br />
René Loser (HP)
Mit HP Technologie bleiben Sie im<br />
Driving Seat<br />
Hr. Patrick Pletscher<br />
Head of Engineering<br />
Senior <strong>Storage</strong> Consultant
• Remote Copy<br />
• SnapShot Technologie<br />
• Recovery Manager for<br />
• VMware<br />
• Microsoft (Exchange, SQL)<br />
• Oracle<br />
• Ausblick, was steht in den nächsten Wochen<br />
an?<br />
Agenda
Metropolitan distance<br />
HP <strong>3PAR</strong> Remote Copy<br />
Daten schützen und bei Bedarf zu Verfügung stellen<br />
Complete<br />
• Fiber Channel or Native IP-based<br />
• No extra copies or infrastructure needed<br />
• Thin provisioning aware<br />
• Thin conversion<br />
• Synchronous, Asynchronous Periodic or<br />
Synchronous Long Distance (SLD)<br />
• Mirror between any <strong>3PAR</strong> array<br />
• Many to one, one to many<br />
• VMware SRM integration<br />
Primary<br />
RC<br />
P<br />
Primary<br />
P<br />
Sync RC<br />
RC1<br />
1:1 Configuration<br />
Sync RC<br />
and / or<br />
Async Periodic RC<br />
Async Periodic RC<br />
Standby<br />
Secondary<br />
P<br />
RC<br />
Synchronous Long Distance<br />
1:2 Configuration<br />
Tertiary<br />
RC2<br />
Secondary<br />
See the demo video on http://h20324.www2.hp.com/SDP/Content/ContentListing.aspx?PortalID=1&booth=66&tag=534&content=3431
HP <strong>3PAR</strong> Remote Copy<br />
N:1 Konfiguration<br />
• You can use Fibre Channel (RCFC) and/or<br />
Remote Copy over IP (RCIP) connections<br />
• Asynchronous Periodic only<br />
• Max support is 4 to 1.<br />
One of the 4 RC relationships can be bidirectionally<br />
• Each RC relationship requires dedicated<br />
node-pairs - In a 4:1 setup the target site<br />
requires 4 node-pairs (8 nodes).<br />
P<br />
Primary Site B<br />
P<br />
Primary Site C<br />
P<br />
Primary Site A<br />
P<br />
RC RC<br />
RC P<br />
RC<br />
Target Site<br />
1:N Konfiguration<br />
• You can use Fibre Channel (RCFC) and/or<br />
Remote Copy over IP (RCIP) connections<br />
• Max support is 1 to 2.<br />
Both mirrors can be bi-directional<br />
• Supported with 1 node pair in primary site<br />
• Supported configurations:<br />
• Async on both target sites<br />
• Sync on 1 site and async on the second site<br />
RC<br />
Primary & Target Site D<br />
P<br />
P<br />
Primary Site<br />
RC<br />
Target Site A<br />
RC<br />
Target Site B
HP <strong>3PAR</strong> Remote Copy<br />
Garantierte Datenintegrität<br />
Single Volume<br />
• All writes to the secondary volume are<br />
completed in the same order as they were<br />
written on the primary volume<br />
Autonomic Multi-Volume Group<br />
• Volumes can be grouped together to maintain<br />
write ordering across sets of volumes<br />
• Useful for databases or other applications that<br />
make dependent writes to more than one<br />
volume<br />
• Secondary groups and volumes are<br />
autonomically created or reconfigured and<br />
credentials inherited<br />
New<br />
Source<br />
Volume<br />
Replicated<br />
Provisioning Group<br />
Primary <strong>3PAR</strong> <strong>Storage</strong><br />
New Target<br />
Volume Created<br />
Autonomically<br />
Replicated<br />
Provisioning Group<br />
Secondary <strong>3PAR</strong> <strong>Storage</strong>
HP <strong>3PAR</strong> Remote Copy<br />
Supportete Transporttechnologien und die maximale Latenzzeit und Distanz<br />
Remote Copy Type<br />
Max Supported Latency<br />
Synchronous FC 2.6ms round trip *<br />
Synchronous IP 2.6ms round trip *<br />
Asynchronous Periodic FC 2.6ms round trip *<br />
Asynchronous Periodic IP<br />
150ms round trip<br />
Asynchronous Periodic FCIP<br />
120ms round trip<br />
*<br />
• Optical Fibre networks typically have a delay of ~5 us/km (0.005ms/km)<br />
• Thus 2.6ms will allow Fibre link distances of up to 260 km<br />
2 x 260km = 520km 520km x 0.005 ms/km = 2.6ms
Peer Persistence v1 – VMware<br />
High Availability Enhancements<br />
What does it provide?<br />
• Load Balancing across data centers with<br />
<strong>Storage</strong> vMotion<br />
• High Availability of VMware environments<br />
across data centers<br />
• Transparent LUN presentation across two sites<br />
within synchronous Remote Copy distances<br />
How does it work?<br />
• Based on <strong>3PAR</strong> Remote Copy and vSphere<br />
ALUA host mode persona 11<br />
• Presents primary Vol as active and secondary<br />
as standby<br />
• Remote Copy Group migration initiated by<br />
“setrcopygroup switchover ”<br />
command<br />
Supported environments:<br />
• ESX vSphere 5.x<br />
• Sync Remote Copy<br />
• Up to RC supported max of 2.6ms (~260km )<br />
Requirements:<br />
• <strong>3PAR</strong> Disk Arrays<br />
• <strong>3PAR</strong> Remote Copy License<br />
• <strong>3PAR</strong> Peer Persistence License<br />
VMware Metro <strong>Storage</strong> Cluster (single subnet)<br />
Fabric A<br />
Fabric B<br />
Up to 2.6ms RTT latency<br />
Vol B sec<br />
Vol B prim<br />
Vol A prim<br />
Vol A sec<br />
Swap Vol A<br />
<strong>3PAR</strong> Array A<br />
<strong>3PAR</strong> Array B
HP <strong>3PAR</strong> Full Copy V1– restorable copy<br />
Bestandteil des <strong>3PAR</strong> Betriebssystems<br />
Base Volume<br />
• Full physical point-in-time copy<br />
• Provisionable after copy ends<br />
• Independent of base volume’s RAID and<br />
physical layout properties<br />
• Fast resynchronization capability<br />
• Thin Provisioning-aware<br />
• Full copies can consume same physical<br />
capacity as thinly provisioned base volume<br />
Full Copy<br />
Intermediate<br />
Snapshot<br />
Full Copy
HP <strong>3PAR</strong> Full Copy V2 – instantly accessible copy<br />
Bestandteil des <strong>3PAR</strong> Betriebssystems<br />
Base Volume<br />
• Share data quickly and easily<br />
• Full physical point-in-time copy<br />
• Immediately provisionable to hosts<br />
• Independent of base volume’s RAID and<br />
physical layout properties<br />
• No resynchronization capability<br />
• Thin Provisioning-aware<br />
• Full copies can consume same physical<br />
capacity as thinly provisioned base<br />
volume<br />
Full Copy<br />
Intermediate<br />
Snapshot<br />
Full Copy
HP <strong>3PAR</strong> Virtual Copy – Snapshot at its best<br />
Smart<br />
• Individually erasable and promotable<br />
• Scheduled creation/deletion<br />
• Consistency groups<br />
Thin<br />
• No reservation, non-duplicative<br />
• Variable QoS<br />
Ready<br />
• Instantaneously readable and/or writeable<br />
• Snapshots of snapshots of …<br />
• Virtual Lock for retention of read-only snaps<br />
• Automated erase option<br />
Integrated<br />
• MS SQL<br />
• MS Exchange<br />
• Oracle<br />
• vSphere<br />
Base Volume<br />
• HP Data Protector<br />
• …<br />
Up to 8192 Snaps per array<br />
100s of Snaps…<br />
…but only one<br />
CoW required<br />
Top 10 Arrays WW as of July 2012<br />
# of Snapshots Model<br />
6559 V800<br />
6172 S800<br />
6156 S800<br />
5138 S800<br />
4666 S400<br />
4482 S800<br />
4341 T800<br />
4295 T800<br />
3991 T400<br />
3871 T800
HP <strong>3PAR</strong> Virtual Copy Relationships<br />
The following shows a complex relationship scenario
Application SW Suite for VMware<br />
Recovery Manager<br />
for VMware<br />
Host Explorer<br />
VASA Provider<br />
• Protects VMware<br />
VMs/Datastores<br />
• Support File Level restore<br />
• Application consistent VM<br />
Virtual Copy<br />
• Integration with Remote Copy<br />
• Built-in Scheduler<br />
• Local/Remote Virtual Copy<br />
management<br />
• Support Peer Persistence<br />
(Transparent failover)<br />
• Collects ESX(i) configuration<br />
and push it to <strong>StoreServ</strong>.<br />
• Schedule based on userdefined<br />
intervals.<br />
• Information can be viewed<br />
from <strong>StoreServ</strong> IMC/CLI.<br />
• Provide interface to vCenter<br />
SMS client.<br />
• <strong>Storage</strong> array discovery<br />
• <strong>Storage</strong> Event and Alert<br />
reporting<br />
• Capability based provisioning
<strong>3PAR</strong> Recovery Manager for VMware<br />
Array-based Snapshots for Rapid Online Recovery<br />
Solution composed of<br />
• <strong>3PAR</strong> Recovery Manager for VMware<br />
• <strong>3PAR</strong> Virtual Copy<br />
• VMware vCenter<br />
Use Cases<br />
• Expedite provisioning of new<br />
virtual machines from VM copies<br />
• Rapid online recovery of files<br />
• Snapshot copies for testing and<br />
development<br />
Benefits<br />
• Hundreds of VM snapshots granular,<br />
rapid online recovery<br />
• Reservation-less, non-duplicative without agents<br />
• vCenter integration – superior ease of use<br />
See also http://h18006.www1.hp.com/storage/software/3par/rms-vsphere/index.html
Recovery Managers for Microsoft<br />
RM MS Exchange and RM MS SQL<br />
• Automatic discovery of Exchange and SQL Servers and their<br />
associated databases<br />
• VSS Integration for application consistent snapshots<br />
• Support for Microsoft® Exchange Server 2003, 2007, and 2010<br />
• Support for Microsoft® SQL Server2005, 2008 , 2012<br />
(Standalone and always on availability group (A<strong>AG</strong>)<br />
• Database verification using Microsoft tools<br />
Built upon <strong>3PAR</strong> Thin virtual copy technology<br />
• Fast point-in-time snapshot backups of Exchange & SQL<br />
databases<br />
• 100’s of copy-on-write snapshots with just-in-time, granular<br />
snapshot space allocation<br />
• Automatic recovery from snapshot<br />
• <strong>3PAR</strong> Remote Copy integration<br />
• Exporting of database back ups to other hosts<br />
Backup integration<br />
• HP DataProtector<br />
• Symantec NetBackup and Backup Exec<br />
See also<br />
http://h18006.www1.hp.com/storage/software/3par/rms-exchange/index.html<br />
http://h18006.www1.hp.com/storage/software/3par/rms-sql/index.html
RMS & RME Architecture<br />
SQL or Exchange<br />
Production Server<br />
RM Client &<br />
Backup Server<br />
1. Off-host Backup<br />
2. Direct Restore from Tape<br />
3. Direct Mount of Snapshot<br />
4. Restore from Snapshot with<br />
file copy restore<br />
9:00<br />
13:00<br />
<strong>3PAR</strong><br />
Production<br />
Volumes<br />
Snapshots<br />
17:00<br />
Optional Library<br />
Tape or D2D
RM SQL Recovery<br />
Possibilities:<br />
• File Copy Restore to a Point-in-Time<br />
• File Copy at a Database Level<br />
• Volume Restore at a Database Level
Possibilities:<br />
• File Copy Restore<br />
• Volume Restore<br />
RM Exchange Recovery
<strong>3PAR</strong> Recovery Manager for Oracle<br />
Allows PIT Copies of Oracle Databases<br />
• Non-disruptive, eliminating production downtime<br />
• Uses <strong>3PAR</strong> Virtual Copy technology<br />
Allows Rapid Recovery of Oracle Databases<br />
• Increases efficiency of recoveries<br />
• Allows Cloning and Exporting of new databases<br />
• Support single datafile or tablespace restore<br />
Integrated High Availability with Disaster Recovery Sites<br />
• Integrated <strong>3PAR</strong> Replication / Remote Copy for Array to Array<br />
DR<br />
Currently supported with<br />
• Oracle 10g and 11g<br />
• RHEL 4 and 5<br />
• OEL 5<br />
• Solaris 10 SPARC<br />
• HP-UX 11.31<br />
Support planned for<br />
• IBM AIX<br />
Supported Backup <strong>Solutions</strong><br />
• Symantec Netbackup, RMAN and Data Protector (work in<br />
progress)<br />
See also http://h18006.www1.hp.com/storage/software/3par/rms-oracle/index.html
Was kommt in den nächsten Wochen?<br />
• Recover Manager für Hyper-V<br />
• Peer Persitence V2<br />
• QoS<br />
• Data at Rest Encryption für 7000 and 10000<br />
• Grössere Disks und SSDs<br />
• Online Import – Erweiterter Support EVA FW, vSphere, Linux,<br />
HP-UX und AIX Support<br />
• OVMS GA Support
Planned June 2013<br />
Peer Persistence v2 – <strong>3PAR</strong> <strong>Storage</strong> & VMware<br />
What does it provide?<br />
• Automated and manually initiated Load Balancing<br />
across data centers with <strong>Storage</strong> vMotion<br />
• High Availability of VMware environments across data<br />
centers<br />
• Transparent LUN presentation across two sites within<br />
synchronous Remote Copy distances<br />
How does it work?<br />
• Based on <strong>3PAR</strong> Remote Copy and vSphere ALUA host<br />
mode persona 11<br />
• Presents primary Vol as active and secondary as<br />
standby<br />
• Manual volume swap initiated by “setrcopygroup<br />
switchover ” command<br />
• Automated volume swap arbitrated by a Quorum<br />
Witness (QW)<br />
Supported environments:<br />
• ESX vSphere 5.x incl. Metro <strong>Storage</strong> Cluster<br />
• Sync Remote Copy<br />
• Up to new RC supported max of 5ms (~500km )<br />
Requirements:<br />
• <strong>3PAR</strong> Disk Arrays<br />
• <strong>3PAR</strong> Remote Copy License<br />
• <strong>3PAR</strong> Peer Persistence License<br />
VMware vSphere Metro <strong>Storage</strong> Cluster (single subnet)<br />
<strong>3PAR</strong> Array<br />
Site A<br />
Vol B sec<br />
Vol A prim<br />
Fabric A<br />
Fabric B<br />
Up to 5ms RTT latency<br />
QW<br />
vSphere<br />
Site C<br />
Vol B prim<br />
Vol A sec<br />
<strong>3PAR</strong> Array<br />
Site B
Coming soon – Priority Optimization (QoS)<br />
Planned June 2013<br />
• Assure tenant/application level QoS<br />
by assigning Virtual Volume Set<br />
performance caps in IOPS and/or<br />
bandwidth (MB/s)<br />
• Will be available for all <strong>StoreServ</strong><br />
model at <strong>3PAR</strong> OS 3.1.2<br />
Tenant 1<br />
shared 20,000 IOPS<br />
Over-commits at 150%<br />
of assigned IOPS<br />
vv_set/Application 1<br />
(10,000 IOPS)<br />
vv_set/Application 2<br />
(10,000 IOPS)<br />
vv_set/Application 3<br />
(10,000 IOPS)<br />
Tenant 2<br />
dedicated 100,000 IOPS<br />
vv_set/Application 4<br />
(100,000 IOPS)
HP <strong>3PAR</strong> Priority Optimization – planned June 2013<br />
Total IO/sec<br />
Total kB/sec<br />
Read IO/sec<br />
Read kB/sec<br />
Write IO/sec<br />
Write kB/sec<br />
# showqos<br />
setqos -bw -io 2000 none 5000k -bw none vvset:3p74-QoS-vvset<br />
Id # showqos Type Name QoS IOLimit KBLimit<br />
Id 1 vvset Type 3p74-QoS-vvset Name<br />
on IOLimit - KBLimit -<br />
-------------------------------------------<br />
1 vvset 3p74-QoS-vvset on 2000 - - 5000 -<br />
-------------------------------------------<br />
1 total<br />
1 total<br />
Priority Optimization Example:<br />
Windows 2012<br />
Virtual Volume - SSD Raid1<br />
Iometer:<br />
• 100 % random<br />
• 4 kB blocksize<br />
• 80/20 Read/Write
Fragen & Antworten
Agenda<br />
14:30 – 14:45 Begrüssung Alfio Lazzari (LAKE)<br />
14:45 – 15:15 Machen Sie Ihre Daten schlank<br />
Holen Sie mehr aus Ihrer Diskkapazität mit der <strong>3PAR</strong> <strong>StoreServ</strong> 7000 Thin Technologie<br />
15:15 – 15:45 Mit HP Technologie bleiben Sie im Driving Seat<br />
Fredy Wismer (LAKE)<br />
Verkürzen Sie die RPO und RTO Zeiten durch den Einsatz von SnapShot & Datenreplizierung Patrick Pletscher (LAKE)<br />
15:45 – 16:00 Ihr persönliches <strong>StoreServ</strong> <strong>Storage</strong> System<br />
Wir ermitteln für Sie die optimale <strong>StoreServ</strong> Konfiguration durch gezielte Datenerhebung<br />
16:00 – 16:30 Daten schnell und sicher gespeichert mit<br />
HP StoreOnce Backup2Disk<br />
16:30 – 17:30 Pause / Apéro<br />
17:30 – 19:30 Kino Start – Iron Man 3<br />
Daniel Estermann (LAKE)<br />
René Loser (HP)
Ihr persönliches <strong>StoreServ</strong> System<br />
Hr. Daniel Estermann<br />
Senior <strong>Storage</strong> Consultant
Agenda<br />
Sizing <strong>StoreServ</strong> <strong>Storage</strong> System<br />
• LAKE Vorgehen<br />
• Datenerfassung<br />
• Unser Vorgehen bei EVA Kunden<br />
• Fragen welche zu klären sind<br />
• Sizing <strong>3PAR</strong> <strong>Storage</strong><br />
• Q & A
LAKE Vorgehen<br />
1. Datenerfassen (Ist Aufnahme)<br />
2. Datenanalysieren und Auswerten<br />
3. Kundenbedürfnisse jetzt und in der Zukunft<br />
4. System sizing<br />
5. Finale Lösung präsentieren
• Was habe ich heute?<br />
Datenerfassung Kapazität<br />
• Blick auf das heutige <strong>Storage</strong> und deren<br />
Auslastung<br />
• Überprüfung der Datenmenge auf OS Basis mit<br />
Ninja Thin<br />
• Was hat der Kunde für ein Daten Wachstum in<br />
den nächsten Jahren?<br />
• Zukünftige Anforderungen und Projekte
Datenerfassung I/O Leistung<br />
• Messen der heutigen <strong>Storage</strong> Auslastung<br />
• HP EVA <strong>Storage</strong> EVAperf<br />
• Third-Party <strong>Storage</strong> Daten logen wenn möglich<br />
• Messen mit Windows Perfmon<br />
• VMware eigenen oder zusätzlichen Management<br />
Tools<br />
• VMware vCenter<br />
• VMware Operations Manager<br />
• Microsoft System Center<br />
• Third-Party Produkt
• EVA Daten Aufzeichnung<br />
Unser Vorgehen bei EVA Kunden<br />
• 7 Tage EVAperf Daten aufzeichnen<br />
• Kapazitätsermittlung auf OS basierend<br />
• Ninja Thin<br />
• EVAperf auswerten<br />
• EVAperf Daten können direkt ins Ninja Tool<br />
importiert werden<br />
• <strong>Storage</strong> Sizing<br />
• Ninja Sizer
Host scanen
Kapazität Übersicht
IO Auswertung EVAperf<br />
EVAperf Daten importieren<br />
Auswertung EVAperf Daten
• Benötige ich Daten Spiegelung?<br />
• Synchron oder Asynchron?<br />
• Über welche Distanz?<br />
• Benötige ich SnapShots?<br />
• Für einen schnellen Restore<br />
• Für einen Backup<br />
Fragen für den Kunden<br />
• Gibt es in der IT Strategie Vorgaben für RTO und<br />
RPO?<br />
• Was benötigt der Kunde in der Zukunft?<br />
• Zukünftige Anforderungen und Projekte
• Kapazität Daten<br />
• RAW und OS Basieren<br />
Ermittelte Informationen<br />
• Daten Typ ( Archiv, File, DB, OS, Virtuelle)<br />
• IO Daten<br />
• IO gemessen Maximal, Durschnitt und 95% Regel<br />
• IO ermittelt über Anzahl der Festplatten und Typ<br />
im bestehenden <strong>Storage</strong><br />
• Read/Write Verhältnis<br />
• Kunden Anforderungen<br />
• Wachstum und zukünftige Anforderungen
• Anzahl der Disk’s und welchen Typ?<br />
• Multi Tiering?<br />
• Adaptive oder Dynamic Optimization<br />
• SAS und SATA<br />
• SSD und SAS<br />
Aufgabenstellung für das Sizing<br />
• SSD, SAS und SATA<br />
• Benötigte Features? Software Komponenten?<br />
• Replication, Dynamic oder Adaptive Optimization etc.<br />
• <strong>3PAR</strong> Startkonfiguration<br />
• 7200-2N<br />
• 7400-2N oder 7400-4N
<strong>3PAR</strong> Sizing
Kapazität TB<br />
I/O Leistung<br />
Wachstum<br />
Anforderungen<br />
Ihr Persöhnliches <strong>3PAR</strong>
Fragen & Antworten