FlashArray: Purity 6.1 - ActiveCluster over Fibre Channel/FC and other Innovations

[ NOTE: machine translation with the help of DeepL translator without additional proofreading and spell checking ]


Great news to start the year - Purity//FA 6.1.0 is GA for customers.


In 2017, Purity 5.0 released ActiveCluster as a synchronous replication solution over TCP/IP. With increasing implementations of the replication feature came requests to also use Fibre Channel as a replication transport method. These requests can now be fulfilled, allowing customers to take advantage of Fibre Channel for ActiveCluster replication.

So ActiveCluster now offers an option for replication: TCP/IP or FC. The deployment and configuration of AC/FC is almost identical as already known.


Major new features with Purity 6.1.0 are:

  • ActiveCluster over Fibre Channel

  • NVMe over Fibre Channel (NVMe-FC)

  • Cloud Block Store for Microsoft Azure

  • i3en support for virtual instances on Cloud Block Store for AWS //V20

  • Support for pausing CBS array on AWS (operational cost optimization)

  • Enhanced ACL (access control) features for FlashArray File Services:

  • SMB: access based enumeration: files/folders can be hidden from clients that do not have sufficient permissions

  • NFS: read-only exports

  • Purity//FA REST API version 2.4

  • Manual offload/restore of snapshots without protection groups

  • Display DMM (DirectMemory Cache) performance metrics


Insight: ActiveCluster over Fibre Channel


In line with the Pure philosophy (performance and simplicity), the following highlights also arise for ActiveCluster over FC:

  • AC/FC uses NVMe as transport protocol (no SCSI)

  • Auto-discovery of peer replication WWNs (does not correspond to auto-zoning)

  • Display of detailed port statistics for all interfaces (real-time and history)

The requirements for using ActiveCluster over Fibre Channel are:

  • Purity 6.1 and newer

  • FlashArray //XR2 or newer (X10 not supported/as of today)

  • Two dedicated FC ports per controller (four ports per chassis), no host IO port can be mixed with replication traffic. A dedicated HBA is not necessarily required

  • Brocade or Cisco FC switches

  • Latency/RTT not greater than 5ms (with release 6.1.0). No limit on distance - latency is limit.

Limitations:

  • FlashArrays/Peers cannot be directly connected. FC switches must exist between the arrays

  • FC switches should use speeds of 16Gb (limitations when using older switches/speed)

  • No mixed operation ActiveCluster IP and FC - only 1 protocol between array.

  • FC replication is not yet supported for Async/ActiveDR or Active-Active-Async.

(Brief overview) Commissioning and configuration:

  • Selection of FC ports for replication

  • Port zoning

  • ActiveCluster peering of the two FlashArrays

Eine X70R3-FC wird im Standard z. B. mit 1x Quad-Port in Slot0 & 1x Dual-Port Slot2 ausgeliefert. Insofern der Kunde die Ports aus Slot0 für Host-IO-Traffic verwendet, bleiben Port fc8 + fc9 für die FC-Replikation übrig. Es wäre jedoch auch möglich fc2 + fc3 für die Replikation zu nutzen. Der definierte Traffic-Typ wird dann für ausgewählte Ports angepasst.

Beispielhafte Port-Verwendung an einer Dual Fabric Infrastruktur

The zoning


In a single fabric infrastructure (not recommended for redundancy reasons), all replication WWNs are packed into one zone (e.g. PURE-AC-REPLICATION). In total 8 WWNs (4 per array).


Common standard is the dual fabric architecture. Each fabric should have a zone with 4 WWNs.


For our example:


SAN A (Fabric 1) e. g. PURE-AC-REPLICATION_SAN_A

FlashArray-1 ct0.fc8

FlashArray-1 ct1.fc8

FlashArray-2 ct0.fc8

FlashArray-2 ct1.fc8


SAN B (Fabric 2) e. g. PURE-AC-REPLICATION_SAN_B

FlashArray-1 ct0.fc9

FlashArray-1 ct1.fc9

FlashArray-2 ct0.fc9

FlashArray-2 ct1.fc9

provided/copyright Pure Storage

Via the GUI (Health > Connections > Array Connections) you can then check via the new view whether the zoning was configured successfully and redundantly. The FC port statistics are also displayed in the GUI (Health > Network).


The ActiveCluster activation


The further setup will then be done together with the support. The supporter will work through a pre-activation checklist/script and then set a tunable to allow the activation. Usually the peer connection (exchange connection key) is then also set up directly.


For further information please refer to the FlashArray Release Notes in the Pure Technical Services Portal: Purity//FA 6.1.x Release Notes or the Purity//FA 6.1.0 FlashArray User Guide.


Others / Enhancements


Otherwise, there were minor improvements: the default password for the pureuser must now be changed immediately upon startup. For existing systems, an alarm is generated after the upgrade if the default password has not yet been changed.


With Purity //FA 6.1, the vSphere HTML Client plugin version 4.4.0 is now also natively provided on the controllers for push installation.


More info - Links


All officially published setting options in the GUI but also CLI can be read via the "on-board" user guides of the Pure Storage systems.


Click on "Help" in the Purity main menu.


The User Guide is structured like the main menu and can be opened downwards. A search function is also integrated - within here you can also search for keywords.

WEB: Pure Storage (Pure1) support portal - Ticket system and support *(requires registered FlashSystems)

PHONE: Pure Storage phone support: GER - (+49) (0)800 7239467; INTERNATIONAL - (+1) 650 7294088

WEB: Pure Storage community

WEB: Pure Storage OFFICIAL blog

The blog lives from your questions, wishes and suggestions...every comment is welcome. I am very grateful for feedback.

803 views0 comments