Files
root 69b0ccb4b8 Fix 30 KB errors: ExaCS storage 3-10x wrong, ECPU transition, IOPS limits, missing shapes
Cross-referenced all 70+ KB files against official Oracle docs. Critical fixes:
- ExaCS: storage was flash cache not usable disk (76.8→190.8 TB QR), added X11M, removed X10M
- ADB-S: OCPU→ECPU migration, 26ai as default, cross-region RTO 15→10 min
- Compute: E5 max 94→126, A2 80→78, added E6/A4/BM shapes, 11 GPU shapes
- Storage: Higher Perf IOPS 35K→50K, UHP 450K→300K, File Storage 25K→50K
- DBCS: E5.Flex max 94→64, storage 80→150 TB, added 26ai
- Networking: FastConnect +400 Gbps, NSG max 120 rules

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-01 23:59:28 -03:00

106 lines
3.5 KiB
YAML

---
last_verified: 2026-03-14
service: Exadata Cloud Service (ExaCS)
category: database
oci_color: "#AA643B"
---
what: Full Oracle Exadata system in OCI. Provides OS access, RAC, full DBA control
with cloud elasticity. Runs the same Exadata software as on-prem.
when_to_use:
- Need OS-level access for custom agents, scripts, or compliance tools
- RAC required for application architecture
- Sustained high OCPU workloads (>128 OCPUs)
- Lift-and-shift from on-prem Exadata with minimal changes
- Need custom database parameters not available in ADB
- Applications that use features incompatible with ADB (e.g., dblinks to non-Oracle)
when_NOT_to_use:
- Small/medium workloads (over-provisioned)
- No DBA team (ExaCS requires database administration)
- Variable workloads that benefit from ADB auto-scaling
- Budget-constrained (high minimum commitment)
limits:
shapes:
- name: "Exadata.Quarter3.100"
base_ocpu: 100
max_ocpu: 100
storage_tb: 149
notes: "Legacy X8 fixed shape — storage is usable disk"
- name: "Exadata.Half3.200"
base_ocpu: 200
max_ocpu: 200
storage_tb: 299
notes: "Legacy X8 fixed shape"
- name: "Exadata.Full3.400"
base_ocpu: 400
max_ocpu: 400
storage_tb: 598
notes: "Legacy X8 fixed shape; current generation is elastic X9M/X11M"
max_vm_clusters: 8
max_databases_per_cluster: "Limited by resources"
max_pdb_per_cdb: 252
features:
os_access: true
rac: true
data_guard: true
cross_region_data_guard: true
custom_db_home: true
custom_patches: true
grid_infrastructure: true
flash_cache: true
smart_scan: true
storage_compression: true
hybrid_columnar_compression: true
ha_dr:
local_ha: "RAC across Exadata nodes, automatic failover"
data_guard:
same_region: "Synchronous, automatic failover with FSFO"
cross_region: "Asynchronous or far-sync, manual or automatic failover"
sizing_rules:
minimum: "Quarter rack (100 OCPUs base)"
scaling: "Can enable additional OCPUs in increments"
storage: "Comes with infrastructure, expandable"
pricing:
model: infrastructure_plus_ocpu
infrastructure_monthly: "Fixed per shape"
enabled_ocpu_per_hour: 0.336 # approximate
byol_ocpu_per_hour: 0.168
notes:
- "Infrastructure cost is fixed per shape"
- "Only enabled OCPUs are billed"
- "BYOL provides ~50% discount on OCPU portion"
- "Storage included with infrastructure"
gotchas:
- id: EXACS-001
issue: "Provisioning takes hours to days"
impact: "Cannot provision on-demand like ADB-S"
recommendation: "Plan capacity well in advance"
- id: EXACS-002
issue: "Minimum quarter rack commitment"
impact: "Significant base cost even at low utilization"
recommendation: "Ensure sustained usage justifies the infrastructure cost"
- id: EXACS-003
issue: "DBA team required for ongoing operations"
impact: "Patching, tuning, and maintenance are customer responsibility"
recommendation: "Ensure DBA team capacity before choosing ExaCS over ADB"
- id: EXACS-004
issue: "Scale-down requires VM cluster modification"
impact: "Not as elastic as ADB-S for variable workloads"
recommendation: "Right-size for steady-state, not peak"
migration_tools:
- name: Oracle Data Guard
use_for: "On-prem Exadata to ExaCS with minimal downtime"
- name: RMAN Restore
use_for: "Backup-based migration with planned downtime"
- name: Zero Downtime Migration (ZDM)
use_for: "Automated physical migration"