Monday, November 19, 2012

NetApp NCDA NS0-154 Exam Cram Notes: Part 1 of 3


0: Definitions

DOT = Data ONTAP
inodes = Used by DOT in an active file system to reference disk blocks
NetApp FAS = NetApp Fabric-Attached Storage system
NAS = a file based storage system that makes data available over the network using NFS and CIFS protocols
NOW = NetApp on the Web
NS0-154 = code for the Data ONTAP 8.0 7-MODE Administrator exam
SAN = a block-based storage system that makes data available over the network using FC, FCoE, and iSCSI protocols
VLD = Virtualized Local Disk : a pre-iSCSI storage protocol NetApp developed several years ago to satisfy the market need for IP-based storage networking solutions for key Windows NT and Windows 2000 applications. It was developed as a proprietary interim solution until an industry standard protocol with Microsoft support became available.
WAFL = Write Anywhere File Layout : designed by NetApp, it is not a file system but provides mechanisms that enable a variety of file systems and technologies that want access to disk blocks.
WORM = Write Once Read Many

1: NCDA Overview

7-Mode:
Data ONTAP 7-mode (or classic mode) allows FAS arrays to be deployed as a local two-node cluster, a geographically spanned MetroCluster, and as a remote distributed FlexCache

Cluster-Mode:
Data ONTAP C-Mode (or cluster mode) expands a NetApp storage cluster from 2 nodes to 24 nodes, increases the features found in 7-mode to include endless scaling, global name spaces, and the complete separation of data and data access from the hardware layer in the form of next generation vFilers

Data ONTAP 8.0 7-Mode Storage Architecture:
system> sysconfig -r
Disks – parity, double-parity, or data disks
RAID group – contains physical disks, and are either RAID4* or RAID-DP**
Plex – contain RAID groups, and an aggregate will normally have one plex***
Aggregate – contain its own plex(es) and provide storage to a volume or volumes
*RAID 4 is similar to RAID 5 except it uses a dedicated parity disk as opposed to distributed parity
**RAID-DP is similar to RAID 6 except it uses two dedicated parity disks as opposed to distributed parity
***SyncMirror mirrored aggregates have two plexes – plex0 and plex1, with plex1 containing a mirror of plex0's data

Flexible Volumes (FlexVol volume):
A flexible volume:
- Is loosely coupled to its container aggregate
- Is managed separately from the aggregate
- Can be created almost instantaneously
- Can be as small as 20 MB
- Limited to aggregate capacity (if guaranteed)
- Can be as large as the volume capacity supported for the storage system (not guaranteed)
- Can increase and decrease while online (resize without disruption)
- Can size in any increment (as small as 4 KB) and size quickly

Aggregates and FlexVol Volumes:
Three possible FlexVol guarantees
- Volume option (default): space is allocated or “taken away” from the aggregate when the volume is created
- File option: space is allocated from the aggregate when certain “space-reserved” files (such as space-reserved LUN) is created
- None (thin provisioning): space is not allocated from the aggregate until used by the file

2: Miscellaneous

CIFS, FCP, and NFS, can coexist on a NetApp storage system.
In an iSCSI or FC configuration, the host's HBA is referred to as the initiator, and the storage system's HBA is referred to as the target.
Masking controls LUN availability to initiators.
Host multi-pathing describes a FC or IP SAN solution that has at least two distinct physical paths to a LUN.
A pre-requisite before configuring a storage system as a member of a Windows active directory domain, is for the time on the storage system to be set to within +/- 5 minutesof the time on the domain controller.
NTFS allows file access based on NT ACLs and SID.
UNIX allows file access based on GID or UID and UNIX permissions.

The three qtree security styles:
i) NTFS
- For CIFS clients, security is handled using Windows NTFS ACLs.
- For NFS clients, the NFS username is mapped to a Windows username which is then associated with a Windows security identifier (SID) and its group. These mapped credentials are used to determine file access based on the NTFS ACL.
ii) UNIX
- Just like UNIX, files and directories have UNIX permissions.
- For CIFS client, the Windows username is mapped to a UNIX username. This mapped account is then used to determine file access based on the UNIX security.
iii) Mixed
- Both NTFS and UNIX security is allowed. A file or directory can have either NTFS ACLs or UNIX permissions.
- For NTFS ACLs and NFS clients, the NFS username is mapped to a Windows username and its associated groups. These mapped credentials are used to determine file access based on the NTFS ACL.
- For UNIX permissions and CIFS clients, the Windows username is mapped to an NFS user. These mapped credentials are used to determine file access based on the UNIX security.
- The default file security style is the style most recently used to set permissions on the file.

DOT 7G to 8.0 7-mode migration: non-traditional aggregates designated as 32-bit aggregates.
Cannot currently (with 8.0) convert a 32-bit aggregate into 64-bit aggregate.
Maximum size of a 32-bit FlexVol = 16 TB.
Maximum size of a 64-bit FlexVol maximum size of the corresponding aggregate.
Maximum number of FlexVols per system = 500 (FAS2040 & FAS3210 = 200)
When two aggregates are rejoined, one is overwritten and any data changed on it after the split is lost.
DOT will warn if a mirrored volume has a failed disk and there are no available spares.
If the spare disk is larger than a failed disk, DOT will use the disk and right-size it.

3: NFS (and exportfs command)

The storage system provides two types of exports:
- Persistent: defined in /etc/exports and persistent across reboots
- Temporary: defined through command and located in memory only

Exportable NFS resources: volume, directory/Qtree, file

Rules for Exporting Resources:
i: Specify complete path name; must begin with /vol prefix
ii: Cannot export /vol – which is not a path name – to a file, directory or volume (export each volume separately)
iii: When export a resource to multiple targets, separate the target names with a colon (:)
iv: Storage system must resolve hostnames using DNS, NIS or /etc/hosts per order in /etc/nsswitch.conf
v: Exporting ancestors and descendants is allowed (exports can access multiple paths on the same volume)
vi: Storage system determines permissions by matching the longest prefix

exportfs
-a : Exports all file system paths specified in the /etc/exports file.
-b : Enables or disables fencing of specific NFS clients from specific file system paths, giving the NFS clients read-only or read-write access, respectively.
-c : Checks whether NFS clients have a specific type of access to a file system path.
-f : Flushes entries from the access cache.
-h : Displays help for all exportfs options.
-i : Ignores the options specified for a file system path in the /etc/exports file.
-o : Specifies one or more export options for a file system path as a comma-delimited list.
-p : Exports a file system path and adds a corresponding export entry to the /etc/exports file.
-q : Displays the export options for a file system path.
-r : Exports all file system paths specified in the /etc/exports file and unexports all file system paths not specified in the /etc/exports file.
-s : Displays the actual file system path corresponding to an exported file system path.
-u : Unexports a file system path.
-v : Specifies that Data ONTAP should be verbose.
-w : Saves exported file system paths and their export options into a file.
-z : Unexports a file system path and removes its export entry from the /etc/exports file.

4: CIFS (and cifs command)

The SMB 2.0 protocol is not supported in the DOT 8.0 release.
In the circumstance that a CIFS and UNIX user name are the same, automatic mapping will occur if the user name authentication is available.
option wafl.default_security_style causes all new volumes to default to NTFS security style.

cifs command examples:
cifs shares -add sharename /vol/vol2/data : shares out the directory /vol/vol2/data under the name "sharename"

5: iSCSI (and iscsi command)

The iscsi command manages the iSCSI service on a filer
iscsi status : returns whether iscsi service is running or not (if licensed)

iSCSI session established when host initiator logs into iSCSI target. An iSCSI session can have one or more connections.

6: FC (and fcp command)

Both storage systems in the active-active configuration must have the same settings (as defined in the /etc/rc file)

The fcp family of commands manages the Fibre Channel Target adapters and the FCP target protocol.
Syntax examples:
fcp show adapter [ -v ] [ adapter ]
If no adapter name is given, information about all adapters are shown (and FCP target(s) on a storage system.)
fcp show initiator [ -v ] [ adapter ]
If no adapter name is given, information about all initiators (including WWPNs of hosts logged into a storage system) connected to all adapters are shown.

No comments:

Post a Comment