ADM 202 – Data Access and Protection
L9-1
Learning Goals
9.1 Describe Snapshots
9.2 Configure and Use Snapshots
L9-2
What Is a Snapshot?
• A view of a source volume at a specific point in time
• Use to:
– Recover from user errors, data corruption, or program errors
– Create static data sets for queries or auditing
Volume A
L9-3
Characteristics of a Snapshot
• Read-only point-in-time image of a volume
• Tiny space penalty
• Very quick (seconds)
• Persists until user-set expiration
Snapshot
L9-4
Where Do Snapshots Go?
• Top level of every volume
– .snapshot directory
– Exists even if empty
• Not visible to ls –ltarh command
• Accessed via NFS or Hadoop shell
L9-5
Snapshot Architecture
1
2
3
4
5
6
7
8
L9-6
Snapshot Architecture
1
Snapshot 1
2
9 3
10 4
5
11
6
12
7
8
L9-7
Snapshot Architecture
1
Snapshot 1
2
9 3
10 4
5
11 Snapshot 2
6
12
7
8
13'
13
14
L9-8
Snapshot Architecture
1
Snapshot 1
2
9 3
10 4
5
11 Snapshot 2
6
12
7
8
9' 13'
13
14 15
16
10'
L9-9
Snapshot Architecture
1
Snapshot 1
2
9 3
10 4
5
11 Snapshot 2
6
12
7
8
9' 13'
13
14 15
16
10'
L9-10
Snapshot Architecture
1
Snapshot 1
2
9 3
10 4
5
11 Snapshot 2
6
12
7
8
9' 13'
13
14 15
16
10'
Snapshot 3
L9-11
Snapshot Architecture
1
Snapshot 1
2
9 3
10 4
5
11 Snapshot 2
6
12
7
8
9' 13'
13
14 15
16
10'
Snapshot 3
L9-12
Snapshot Architecture
/myvol/file1
1
Snapshot 1
2
9 3
10 4
5
11 Snapshot 2
6
12
7
8
9' 13'
13
14 15
16
10’
Snapshot 3
L9-13
Use Case: Protection From Human Error
Situation:
You know human error is one of the most common
causes of lost or corrupted data, and you want to
protect against loss of time and data.
L9-14
Use Case: Protection From Human Error
Situation:
You know human error is one of the most common
causes of lost or corrupted data, and you want to
protect against loss of time and data.
Solution:
Make frequent snapshots of key data. To control space requirements, set
reasonable retain times.
L9-15
Learning Goals
9.1 Describe Snapshots
9.2 Configure and Use Snapshots
L9-16
Create a Snapshot
L9-17
Create a Snapshot
• Command line:
$ maprcli volume snapshot create –volume <vol name>
-snapshotname <snapshot name>
• Manual snapshots do not expire automatically
Snapshot
L9-18
Scheduled Snapshots
• Scheduled snapshots automatically expire
• 3 pre-defined schedules:
– Normal data
L9-19
Scheduled Snapshots
• Scheduled snapshots automatically expire
• 3 pre-defined schedules:
– Normal data
– Important data
L9-20
Scheduled Snapshots
• Scheduled snapshots automatically expire
• 3 pre-defined schedules:
– Normal data
– Important data
– Critical data
L9-21
Custom Schedules
$ maprcli schedule create –schedule '<JSON object>'
L9-22
Schedule a Snapshot
$ maprcli volume modify –name <vol> -schedule <schedule ID>
L9-23
Display Schedules in MCS
L9-24
Recover Data From a Snapshot
[root@ip-10-0-0-41 cobra-dev]# hadoop fs –mkdir /recovered-files
L9-25
Recover Data From a Snapshot
[root@ip-10-0-0-41 cobra-dev]# hadoop fs –cp /projects/rnd/cobra/cobra-dev.snapshot/2015-10-
14.08-00-00/* /recovered files
L9-26
Recover Data From a Snapshot
Use Linux commands, if the cluster file system is mounted
L9-27
Preserve a Snapshot
Preserving a snapshot removes its expiration date
L9-28
Maintain Snapshots
• Scheduled snapshots are deleted automatically
• Manual snapshots must be deleted manually (through the CLI or MCS)
Keep on top of your manual snapshots!
L9-29
Lab 9.2: Mapr – Snapshot
1_Lab Seperator
L9-30
Next Steps
ADM 202 – Data Access and Protection
L9-31
