Teradata Data Protection

LOCKS
1. Types
  1. Exclusive Locks
    1. Only to databases and tables
    2. Never to rows
    3. Used for DDL statements
    4. Ex: CREATE TABLE
  2. Write Locks
    1. Used when update without where clause
    2. No new write or read locks allowed till write locks released
    3. EX: UPDATE
  3. Read Locks
    1. Several users may hold
    2. No modification of data is permitted
    3. Ex: SELECT
  4. Access Locks
    1. Allows reading data while modification is going on
    2. This lock is for who is not concerned about data integrity
    3. Designed for decision support while reading large table which are updated only by small changes and single row
    4. Stale locks since it reads stale data
2. Levels
  1. Locks can be applied to following levels
    1. Database
    2. Table/View
    3. Locks all row with same row hash
      1. Primary, Fallback row and secondary subtable rows
3. Maintain data integrity and concurrenty controls
4. User can upgrade and downgrade locks
Transient Journal
1. Called as transaction log in other database
2. maintains a copy on each AMP of before images of all rows affected by the transaction
3. Automatic and transparent to the user
4. Provides rollback in the event of TXN failure.
5. Before images are reapplied to table if TXN fails.
6. Before images are discarded upon TXN completion.
RAID Protection
1. Protection Schemes
  1. Raid 1 (Disk Mirroring)
    1. NCR Recommended
    2. Each physical disk in the array has an exact copy in the same array.
    3. The array controller can read from either disk and write to both.
    4. Good Performance with disk failure
    5. High cost in terms of disk space(Mirroring reduces available disk space by 50%.)
    6. Array controller reconstructs failed disks quickly.
  2. Raid 5 (Parity Checking)
    1. Reduced performance with disk failure
    2. Low cost interms of disk space
    3. Data and parity are striped across a rank of 4 disks.(Parity reduces available disk space by 25% in a 4-disk rank.)
    4. If a disk fails, any missing block may be reconstructed using the other three disks.
    5. Array controller reconstruction of failed disks takes longer than RAID 1.
  3. Raid S (Same as Parity checking but EMC disk array used)
FALLBACK
1. Fallback Table(A Fallback table is fully available in the event of an unavailable AMP.)
2. Fallback Row(A Fallback row is a copy of a primary row stored on a different AMP in the same cluster.)
3. Benefits
  1. Permits access to table data during AMP off-line period
  2. Adds a level of data protection beyond disk array RAID 1 & 5
  3. Automatically restores data changed during AMP off-line
  4. Critical for high availability applications
  5. Highest level of data protection is RAID 1 and Fallback
4. Costs
  1. Twice the disk space for table storage is needed
  2. Twice the I/O for INSERTs, UPDATEs and DELETEs is needed
FALLBACK CLUSTER
1. Group of AMPs which act as fallback unit
2. Size: 2 to 16 AMPs per cluser
  1. The larger the size of the cluster, the less noticeable the workload increase is within that cluster when one AMP fails.
  2. 33% of workload increas when 1 AMP failes in the cluster
  3. 2 AMPs failed in same cluster RDBMS halts
  4. Teradata recommends 4 AMPs per cluster
3. Fallback rows for AMPs in a cluster reside in the cluster.
4. Loss of an AMP in the cluster permits continued table access.
5. Loss of two AMPs in the cluster causes the RDBMS to halt.
6. System performance IS adversely affected when one AMP has a disproportionate burden.
Down AMP Recovery Journal (DARJ)
1. IS
  1. Automatically activated when an AMP is taken off-line
  2. Maintained by other AMPs in the cluster
  3. Totally transparent to users of the system
2. While AMP is offline
  1. Journal is active
  2. Table updates continue as normal
  3. Journal logs Row-IDs of changed rows for down-AMP
3. When AMP is back on-line
  1. Restores rows on recovered AMP to current status
  2. Journal discarded when recovery complete
CLIQUES
1. Two or more TPA nodes having access to the same disk arrays are called a clique.
  1. A clique protects you from node failure.
  2. AMP vprocs can run on any node within the clique and still have full access to their disk array space. If a node fails, AMPs migrate to another node in the clique.
2. PEs for channel-attached connections cannot migrate during a node failure, because they depend on the channel hardware physically attached to their node.
3. PEs for lan attached do migrate during a node failure
Archiving and recovering Data
1. ARC (Archive recovery Utility)
  1. Dump database objects for backup or disaster recovery
  2. Subtopic
  3. Restore tables after corruption from failed batch processes
  4. Recover accidently dropper tables, views or macros
  5. Recover from miscellaneous user errors
  6. can copy a table and restore it to another Teradata database
2. BakBone NetValut
3. Veritas Netbackup
PERMANENT JOURNALS
1. An optional, user-specified, system-maintained journal used for database recovery to a specified point in time.
  1. Used for recovery from unexpected hardware or software disasters.
  2. May be specified for: – One or more tables – One or more databases
  3. Permits capture of before images for database rollback.
  4. Permits capture of after images for database rollforward.
  5. Permits archiving change images during table maintenance.
  6. Reduces need for full-table backups.
  7. Provides a means of recovering NO FALLBACK tables.
  8. Requires additional disk space for change images.
  9. Requires user intervention for archive and recovery activity.
2. Before Journal
3. After Journal
4. Dual Before Journal
5. Dual after journal
QUIZ
1. Subtopic