Data write request process in Netapp Clustered Data ONTAP

For SEO: data write request process

Clustered ONTAP system can be managed either from CLI or GUI (system manager or element manager). There are three different shells available with different scopes: Clustershell, nodeshell, systemshell

Step:1 Access through CLI

Clustershell:
———

-SSH is default method
-One can access clustershell using cluster management LIF (recommended) or the node management LIFs
-Scope: entire cluster

LIFs Management: ‘net int show‘ with display the output of available cluster management and node management LIFs along with their IP’s, current node, current port and status

Nodeshell:
——–

-Access nodeshell within the clustershell
-Scope: limited to one node at a time
-Command to access nodeshell is:

Clus::> system node run

Systemshell:
———

-Access systemshell from within the clustershell
-This is to access BSD shell
-Scope: single node at a time
-To access systemshell:

a. security login unlock -username diag
b. system node systemshell
c. Login as “diag”

Step: 2 Access through GUI

-Start the system manager client using the IP address of the cluster management LIF
-This is well known interface to all the existing 7-mode administrators and I don’t think more explanation is needed here

Dataflow in NetApp cluster-mode

To understand the dataflow in cluster-mode, one needs to understand the four major software components along with other key software components on every node in the cluster, they are as follows:

a. N-blade, network “blade”
b. D-blade, data “blade”
c. SCSI-blase, SCSI SAN “blade”
d. M-host, Management blade

Other components:
————–

-The cluster session manager (CSM)
-RDB units, which is the volume location database

All the above mentioned components are accessed only by well-defined application APIs. The N-blade and SCSI-blade translates client requests into Spin Network Protocol (SpinNP) requests. The D-blade contains the WAFL file system that handles the SpinNP requests. The CSM is the SpinNP layer between the N-blade, the SCSI-blase and the D-blade.

Key Concepts:
————

N-blade:  The N-blade handles the networking, NFS and CIFS requests and then translates to SpinNP requests as inputs to CSM

SCSI-blade: The SCSI-blade handles the networking, FC, FCoE and iSCSI requests and then translates to SpinNP requests as inputs to CSM

CSM: CSM then acts as a communicator between N-bade or SCSI-blade SpinNP requests and input them to the D-blade over UDP/IP

D-blade: The D-blade which contains the WAFL file systems then handles the SpinNP requests and them communicate disks and tape devices using FC

Dataflow on a local node:
——————-

1. NAS or SAN clients sends the write requests to a data logical interface (LIF)
2. N-blade (NAS) or SCSI-blade (SAN) which are associated with that LIF translates the NFS or CIFS, FC, FCoE, or iSCSI request to a SpinNP request
3. The SpinNP requests then passed to the D-blade via CSM
4. D-blade then sends the data to NVRAM and to the disks
5. The acknowledgments then works its way back to the client

Dataflow on a Remote node:
——————————————-

1. NAS or SAN clients sends the write requests to a data logical interface (LIF)
2. N-blade (NAS) or SCSI-blade (SAN) which are associated with that LIF translates the NFS or CIFS, FC, FCoE, or iSCSI request to a SpinNP request
3. The SpinNP requests then passed to the “remote” D-blade via Cluster session manager.
4. D-blade then sends the data to NVRAM, NVRAM has two parts one part holds the data which 50% of the NVRAM and time period is 10secs. Whenever first part 50% is fulled it will send to disks for write, either after 10secs it will automatically sends to disks for write when 50% is not full.
5. The acknowledgments then works its way back to the client

data write request process

NCSA Certification Guide Don’t miss

Netapp how to’s