Tiered Storage
Tacnode supports configurable tiered storage across different media to meet business requirements for cost optimization and performance. This guide covers implementing and managing tiered storage strategies for your database workloads.
Overview
Tiered storage allows you to optimize costs and performance by placing data on appropriate storage media based on access patterns and business requirements.
Storage Tiers Available
Storage Tier | Location | Media Type | Use Case |
---|---|---|---|
Hot Storage | /hot (default) | High-Performance SSD | Frequently accessed data requiring low latency |
Cold Storage | /cold | HDD | Infrequently accessed data prioritizing cost savings |
Tablespace Concept
A tablespace is a logical storage zone that defines the physical storage location of data. In Tacnode:
- Tables, indexes, or entire databases can be bound to designated tablespaces
- Tablespaces can be assigned during creation or modification
- Each tablespace maps to a specific storage location
Cold Storage Performance Optimization
When using Cold storage in a Nodegroup, bind a cache to improve access performance:
- Create Cache: In Console → "Data" → "Cache", create new cache or reuse existing resources
- Bind Cache: Associate the cache with your Nodegroup to accelerate cold data access
Tablespace Management
Creating Tablespaces
Create tablespaces to define storage locations for your data:
Parameters:
tablespace_name
: Unique name for the tablespaceOWNER
: Optional owner specificationlocation
: Storage tier location (/hot
or/cold
)
Examples:
Privileges Required
Creating tablespaces requires Tacnode superuser privileges.
Dropping Tablespaces
Remove unused tablespaces when no longer needed:
Examples:
Prerequisites
Ensure no tables or indexes are using the tablespace before dropping it.
Assigning Data to Tablespaces
Table Creation with Tablespace
Specify tablespace during table creation:
Index Creation with Tablespace
Place indexes strategically based on usage patterns:
Migrating Existing Data
Move existing tables and indexes between tablespaces:
Migration Impact
Changing tablespace rewrites all data. During this operation:
- Storage usage temporarily doubles until completion
- The operation may take significant time for large tables
- Consider running during maintenance windows
Database-Level Configuration
Setting Default Tablespace
Configure default tablespaces for new objects:
Per-Session Configuration
Configure tablespace settings for specific sessions:
Partitioned Table Tiering Strategies
Time-Based Tiering
Implement automatic tiering based on data age:
Usage-Based Tiering
Tier data based on access patterns:
Monitoring and Optimization
Checking Tablespace Usage
Monitor space utilization across tablespaces:
Performance Monitoring
Track performance metrics by storage tier:
Best Practices
Data Classification Strategy
-
Hot Storage Candidates:
- Current transactional data
- Frequently queried reference tables
- Real-time analytics data
- Recent log data
-
Cold Storage Candidates:
- Historical archive data
- Backup and recovery data
- Compliance and audit logs
- Infrequently accessed reference data
Optimization Guidelines
-
Cache Strategy:
-
Index Strategy:
- Place frequently used indexes in hot storage
- Use partial indexes to reduce cold storage footprint
- Consider index-only scans for cold data
-
Monitoring Strategy:
- Set up alerts for tablespace utilization
- Monitor query performance across tiers
- Track data access patterns for tiering decisions
Migration Planning
-
Assessment Phase:
- Analyze current data access patterns
- Identify candidates for each storage tier
- Estimate storage and performance impact
-
Implementation Phase:
- Start with non-critical data
- Plan migration during low-traffic periods
- Monitor performance impact
-
Validation Phase:
- Verify query performance meets requirements
- Confirm storage cost objectives achieved
- Document lessons learned for future migrations
Gradual Migration
Consider implementing tiered storage gradually:
- Start with clear archive data candidates
- Monitor performance and cost impact
- Expand tiering strategy based on results
- Automate tier transitions where possible
Troubleshooting
Common Issues
Issue: High latency on cold storage queries
- Solution: Implement caching strategy or move frequently accessed data to hot storage
Issue: Tablespace migration taking too long
- Solution: Break large migrations into smaller chunks or schedule during maintenance windows
Issue: Running out of space in hot storage
- Solution: Review data access patterns and migrate appropriate data to cold storage
Performance Optimization
This tiered storage strategy enables you to optimize both cost and performance by placing data on appropriate storage media based on business requirements and access patterns.