Data Flashcards
1
Q
List supported data reduction techniques
A
- Compression
- Deduplication
- Erasure coding
2
Q
Type of compressions
A
- Inline
- Post process
3
Q
Compression considerations
A
- Configured at container level
- Can be switched on the fly
4
Q
Dedup
A
- Container level
- Based on inline finger prints
- Distributed across all nodes
- can be switched on the fly
5
Q
Erasure coding
A
- Only apply to cold data not accessed for 7 days
- Minimum of 4 nodes
- Configured at container level
- can be switched on the fly
6
Q
What are hybride nodes?
A
Nodes with both SSD and HDD
7
Q
Cluster Capacity
A
HDD+SDD
8
Q
Tier Storage
A
Data is written to SSD and flush to HDD
9
Q
Disk balancing
A
- balancing data across the nodes
- avoid capacity utilisation imbalance
- curator framework
10
Q
Nutanix Snapshots
A
- Redirect on write
- Improved performance with snapshopt
- Replication is efficient and built-in
- Store locally, replicated remotely
11
Q
Types of replications and DR options?
A
- Async
- Near sync
- Sync
12
Q
Async replication
A
- RPO 60 min
- All hypervisors
- All licenses
13
Q
Sync replication
A
- RPO 0
- ESXi & Hyper-V
- Ultimate licenses
14
Q
Near Sync replication
A
- RPO 1 min
- All hypervisors
- Ultimate license
15
Q
CVM Unavailable
A
1) IO is redirected to CVMs other nodes
2) VM is neither moved nor restarted
16
Q
Disk Failure
A
1) Copy the impacted data from other nodes
2) Start self-healing
17
Q
Node Failure
A
1) VM is rebooted on another node
2) Impacted data copied from other nodes
2) Start self-healing
18
Q
Resilience and Self Healing
A
- Nutanix assumes eventual HW failure
- RF2 is common
- Watch out for warnings
- The larger the cluster the faster the rebuild
