Dell PowerProtect DD Deploy 2023 cheat sheet

Data Domain systems run the DD OS on dedicated appliance hardware; the Data Management Engine (DME) handles file system metadata while the Data Movement Engine streams data to disk.

Deduplication is inline and variable-length: the system segments the stream, hashes segments, and stores only unique segments in containers; the index (segment fingerprint lookup) is kept in memory and on disk.

Compression runs after deduplication; local compression plus optional global compression reduce the unique-segment footprint, so effective capacity = logical data / (dedup ratio x compression ratio).

The Data Invulnerability Architecture includes checksums on every block, write verification, and continuous fault detection with background verification (scrubbing) of stored data.

DD3300, DD6300, DD6800, DD9300, DD9500, DD9800 and the DD6900/DD9400/DD9900 generations differ in CPU, memory, and maximum shelf count; capacity scales by adding external ES/DS shelves.

Cache tiers: NVRAM holds the write cache and metadata journal for power-loss protection; SSDs (or cache devices) accelerate metadata and index lookups on larger models.

Front-end connectivity options include 10/25/40/100 GbE Ethernet and Fibre Channel HBAs; back-end SAS or NVMe paths connect controller to disk shelves.

The DD System Manager GUI, the DD CLI (SSH), and REST API are the three administrative interfaces; the CLI is the authoritative path for most configuration and troubleshooting tasks.

High availability pairs (DD9800/DD9900-class) use active-passive controllers with failover; single-node systems rely on replication and backup copies for availability.

Sizing inputs: initial logical data volume, expected daily change rate, retention period, required dedup ratio, and growth forecast; usable capacity must cover full backups plus retention of changed data.

Rule of thumb for capacity: usable capacity >= (full backup size x number of retained fulls) / expected dedup ratio, then apply a headroom margin (commonly 20-30 percent) for metadata and scrub space.

Rack planning: verify rack depth and rail kit compatibility, weight per shelf with drives populated, and floor loading; a fully populated high-end shelf can exceed 60 kg.

Power: provide redundant PDUs on separate feeds, confirm voltage/frequency (200-240 V for most high-end models), and calculate circuit load from the model's published power draw at expected load.

Cooling and airflow: front-to-back airflow requires cold aisle at the front; maintain clearance and ambient temperature within the published operating range (typically 10-35 C).

Network planning: reserve separate VLANs/subnets for management, backup data, and replication traffic; replication links need bandwidth >= change rate per replication window.

Cabling: document Ethernet/FC port mapping, use supported optics and cable lengths, and label both ends; back-end shelf cabling must follow the model's supported topology.

Site readiness checklist: DNS/NTP reachability, IP addresses and gateways, firewall ports (for example 22, 443, 2049, 2052, 111, 3009 for replication), and licensed feature keys on hand.

First boot uses the serial console or the initial GUI wizard to set the system hostname, IP address, netmask, gateway, DNS servers, and admin password.

The sysadmin account is the default full-privilege account; create role-based accounts (admin, user, backup-operator, security-officer) rather than sharing sysadmin.

Set the system time source with NTP (ntp add server then enable ntp); Kerberos, AD, and replication depend on clock accuracy.

License installation: apply the license file with license add or via GUI; verify with license show; features include replication, encryption, extended retention, and DD Boost.

Configure the system identity and location with system set and set the admin email/alerting so SNMP traps and autosupport emails reach operations.

Enable autosupport (support autosupport notify) and configure the mail server; this drives proactive case creation and weekly reporting.

Run config setup (interactive wizard) or configure settings individually; verify with config show and net show, then commit with net apply.

Security baseline: change default passwords, disable unused services, enable FIPS mode if required, and configure the firewall (adminaccess) to restrict management subnets.

Interfaces are configured with net config [] and activated with net apply; verify with net show configuration and net show stats.

Bonding/link aggregation: use LACP (mode 4) for switch-supported aggregation; balance-alb (mode 6) works without switch configuration but is less common in production.

VLAN tagging: create a VLAN interface with net create interface vlan. and attach to the physical parent; multiple VLANs can share one physical link.

Replication traffic should use dedicated interfaces; configure replication with replication add and set the replication port (default 2051 for DD Replicator control, data on 2052).

Routing: add static routes with net route add / ; the default gateway is set per interface or via net route add default.

DNS: net set dns ; verify resolution with net lookup ; DNS failure breaks CIFS/AD joins and replication by hostname.

Firewall: adminaccess add restricts management access; verify allowed subnets with adminaccess show to avoid lockout.

Troubleshooting tools: net ping, net traceroute, net lookup, and net show stats for errors/drops; check interface status with net show hardware.

Storage is organized into the active tier (dedup storage on disk), optional archive tier, and cloud tier (for supported models), enabling tiering of cold data to object storage.

External shelves are added with storage add enclosure ; verify with storage show all and check that the shelf is licensed and detected.

Filesystem capacity is reported by filesys show space; the key figure is the percentage of usable capacity consumed, and the system enters read-only mode at 100 percent (or when the threshold is reached).

Licensing model: capacity-based licenses (per TB of usable or raw capacity) plus feature licenses; verify installed licenses with license show and check expiration.

Dedup ratio = logical (pre-dedup) data / physical (post-dedup) data; global compression adds to this, so the effective ratio reported by filesys show compression combines both.

Cleanup and garbage collection: filesys clean runs reclamation of unreferenced segments; schedule during low I/O; space is not returned until cleanup runs.

Capacity alarms: configure alerts at thresholds (for example 70/85/95 percent) via alerts add; monitor with filesys show space and system show performance.

Cloud tier requires a validated object store and a cloud unit license; movement is policy-driven (data movement policy) and data is rehydrated on recall.

DD Boost distributes part of the dedup work to the backup server (client-side dedup), reducing network traffic; it uses port 2049 (and 2051/2052 for replication) and requires a DD Boost license.

DD Boost supports Fibre Channel (DD Boost over FC) and IP; configure storage units (ddboost storage-unit create) to isolate clients and set quotas.

NFS exports are created with nfs add / [options]; verify with nfs show; NFSv3 and NFSv4 are supported depending on DD OS version.

CIFS/SMB shares are created with cifs share create path and require the system to join AD with cifs setup or authentication commands.

VTL (virtual tape library) presents emulated tape libraries and drives over FC; create with vtl add library and map drives to initiators.

NDMP supports tape-less backup to DD and three-way NDMP; enable with ndmp and configure users and port 10000.

Client integration: install the DD Boost plug-in or use native protocols (NFS/CIFS) with the backup application; verify connectivity with ddboost show connections and status.

Access control: define clients with ddboost clients add and set per-client authentication; restrict NFS exports to specific subnets and avoid no_root_squash unless required.

DD Replicator is a licensed feature; configure with replication add and verify with replication show config.

Replication can be directory-level (MTree), collection-based, or full-system; MTree replication replicates only the specified directory tree and is the most common for backup data.

Replication pairs are defined source-to-destination; initialize with replication initialize and monitor with replication show status and replication show stats.

Bandwidth throttling: set with replication throttle and schedule windows so replication does not saturate backup links; recovery point objective drives the required bandwidth.

RPO calculation: bandwidth >= (daily change rate) / (replication window in seconds); the initial baseline copy transfers full data and may need a longer window.

Replication topology: one-to-one, one-to-many (fan-out), many-to-one, and cascaded; choose based on DR site count and link capacity.

Failover: promote the destination by breaking the replication context (replication break) and reconfiguring clients; plan the reverse replication after recovery.

Security: replication traffic can be encrypted with replication encryption; verify with replication show config and confirm ports 2051/2052 are open between sites.

Health checks: system show hardware, system show alerts, filesys show space, and alerts show current give a quick status snapshot.

Performance: system show performance, net show stats, and disk show stats reveal bottlenecks; check CPU, memory, and disk utilization together.

Logs: review /ddvar/log/messages, ddfs.info, and the autosupport bundle; generate a support bundle with support bundle generate for escalation.

Common issues: replication lag (check link bandwidth and throttle), space full (run filesys clean and review retention), and authentication failures (verify NTP and AD/Kerberos).

Alerts: configure alert notifications with alerts add and verify with alerts show; SNMP traps can be sent to a monitoring platform.

Verification of backups: use ddboost show stats and the backup application's verification; confirm dedup ratio and throughput trends over time.

Upgrades: check the DD OS release notes for supported upgrade paths; use system upgrade and verify with system show version before and after.

Troubleshooting methodology: isolate layer (physical link, IP, service port, application), confirm with ping/traceroute and port checks, then review logs and escalate with a support bundle.

Data Domain systems run the DD OS on dedicated appliance hardware; the Data Management Engine (DME) handles file system metadata while the Data Movement Engine streams data to disk.

Deduplication is inline and variable-length: the system segments the stream, hashes segments, and stores only unique segments in containers; the index (segment fingerprint lookup) is kept in memory and on disk.

Compression runs after deduplication; local compression plus optional global compression reduce the unique-segment footprint, so effective capacity = logical data / (dedup ratio x compression ratio).

The Data Invulnerability Architecture includes checksums on every block, write verification, and continuous fault detection with background verification (scrubbing) of stored data.

DD3300, DD6300, DD6800, DD9300, DD9500, DD9800 and the DD6900/DD9400/DD9900 generations differ in CPU, memory, and maximum shelf count; capacity scales by adding external ES/DS shelves.

Cache tiers: NVRAM holds the write cache and metadata journal for power-loss protection; SSDs (or cache devices) accelerate metadata and index lookups on larger models.

Front-end connectivity options include 10/25/40/100 GbE Ethernet and Fibre Channel HBAs; back-end SAS or NVMe paths connect controller to disk shelves.

The DD System Manager GUI, the DD CLI (SSH), and REST API are the three administrative interfaces; the CLI is the authoritative path for most configuration and troubleshooting tasks.

High availability pairs (DD9800/DD9900-class) use active-passive controllers with failover; single-node systems rely on replication and backup copies for availability.

Sizing inputs: initial logical data volume, expected daily change rate, retention period, required dedup ratio, and growth forecast; usable capacity must cover full backups plus retention of changed data.

Rule of thumb for capacity: usable capacity >= (full backup size x number of retained fulls) / expected dedup ratio, then apply a headroom margin (commonly 20-30 percent) for metadata and scrub space.

Rack planning: verify rack depth and rail kit compatibility, weight per shelf with drives populated, and floor loading; a fully populated high-end shelf can exceed 60 kg.

Power: provide redundant PDUs on separate feeds, confirm voltage/frequency (200-240 V for most high-end models), and calculate circuit load from the model's published power draw at expected load.

Cooling and airflow: front-to-back airflow requires cold aisle at the front; maintain clearance and ambient temperature within the published operating range (typically 10-35 C).

Network planning: reserve separate VLANs/subnets for management, backup data, and replication traffic; replication links need bandwidth >= change rate per replication window.

Cabling: document Ethernet/FC port mapping, use supported optics and cable lengths, and label both ends; back-end shelf cabling must follow the model's supported topology.

Site readiness checklist: DNS/NTP reachability, IP addresses and gateways, firewall ports (for example 22, 443, 2049, 2052, 111, 3009 for replication), and licensed feature keys on hand.

First boot uses the serial console or the initial GUI wizard to set the system hostname, IP address, netmask, gateway, DNS servers, and admin password.

The sysadmin account is the default full-privilege account; create role-based accounts (admin, user, backup-operator, security-officer) rather than sharing sysadmin.

Set the system time source with NTP (ntp add server then enable ntp); Kerberos, AD, and replication depend on clock accuracy.

License installation: apply the license file with license add or via GUI; verify with license show; features include replication, encryption, extended retention, and DD Boost.

Configure the system identity and location with system set and set the admin email/alerting so SNMP traps and autosupport emails reach operations.

Enable autosupport (support autosupport notify) and configure the mail server; this drives proactive case creation and weekly reporting.

Run config setup (interactive wizard) or configure settings individually; verify with config show and net show, then commit with net apply.

Security baseline: change default passwords, disable unused services, enable FIPS mode if required, and configure the firewall (adminaccess) to restrict management subnets.

Interfaces are configured with net config [] and activated with net apply; verify with net show configuration and net show stats.

Bonding/link aggregation: use LACP (mode 4) for switch-supported aggregation; balance-alb (mode 6) works without switch configuration but is less common in production.

VLAN tagging: create a VLAN interface with net create interface vlan. and attach to the physical parent; multiple VLANs can share one physical link.

Replication traffic should use dedicated interfaces; configure replication with replication add and set the replication port (default 2051 for DD Replicator control, data on 2052).

Routing: add static routes with net route add / ; the default gateway is set per interface or via net route add default.

DNS: net set dns ; verify resolution with net lookup ; DNS failure breaks CIFS/AD joins and replication by hostname.

Firewall: adminaccess add restricts management access; verify allowed subnets with adminaccess show to avoid lockout.

Troubleshooting tools: net ping, net traceroute, net lookup, and net show stats for errors/drops; check interface status with net show hardware.

Storage is organized into the active tier (dedup storage on disk), optional archive tier, and cloud tier (for supported models), enabling tiering of cold data to object storage.

External shelves are added with storage add enclosure ; verify with storage show all and check that the shelf is licensed and detected.

Filesystem capacity is reported by filesys show space; the key figure is the percentage of usable capacity consumed, and the system enters read-only mode at 100 percent (or when the threshold is reached).

Licensing model: capacity-based licenses (per TB of usable or raw capacity) plus feature licenses; verify installed licenses with license show and check expiration.

Dedup ratio = logical (pre-dedup) data / physical (post-dedup) data; global compression adds to this, so the effective ratio reported by filesys show compression combines both.

Cleanup and garbage collection: filesys clean runs reclamation of unreferenced segments; schedule during low I/O; space is not returned until cleanup runs.

Capacity alarms: configure alerts at thresholds (for example 70/85/95 percent) via alerts add; monitor with filesys show space and system show performance.

Cloud tier requires a validated object store and a cloud unit license; movement is policy-driven (data movement policy) and data is rehydrated on recall.

DD Boost distributes part of the dedup work to the backup server (client-side dedup), reducing network traffic; it uses port 2049 (and 2051/2052 for replication) and requires a DD Boost license.

DD Boost supports Fibre Channel (DD Boost over FC) and IP; configure storage units (ddboost storage-unit create) to isolate clients and set quotas.

NFS exports are created with nfs add / [options]; verify with nfs show; NFSv3 and NFSv4 are supported depending on DD OS version.

CIFS/SMB shares are created with cifs share create path and require the system to join AD with cifs setup or authentication commands.

VTL (virtual tape library) presents emulated tape libraries and drives over FC; create with vtl add library and map drives to initiators.

NDMP supports tape-less backup to DD and three-way NDMP; enable with ndmp and configure users and port 10000.

Client integration: install the DD Boost plug-in or use native protocols (NFS/CIFS) with the backup application; verify connectivity with ddboost show connections and status.

Access control: define clients with ddboost clients add and set per-client authentication; restrict NFS exports to specific subnets and avoid no_root_squash unless required.

DD Replicator is a licensed feature; configure with replication add and verify with replication show config.

Replication can be directory-level (MTree), collection-based, or full-system; MTree replication replicates only the specified directory tree and is the most common for backup data.

Replication pairs are defined source-to-destination; initialize with replication initialize and monitor with replication show status and replication show stats.

Bandwidth throttling: set with replication throttle and schedule windows so replication does not saturate backup links; recovery point objective drives the required bandwidth.

RPO calculation: bandwidth >= (daily change rate) / (replication window in seconds); the initial baseline copy transfers full data and may need a longer window.

Replication topology: one-to-one, one-to-many (fan-out), many-to-one, and cascaded; choose based on DR site count and link capacity.

Failover: promote the destination by breaking the replication context (replication break) and reconfiguring clients; plan the reverse replication after recovery.

Security: replication traffic can be encrypted with replication encryption; verify with replication show config and confirm ports 2051/2052 are open between sites.

Health checks: system show hardware, system show alerts, filesys show space, and alerts show current give a quick status snapshot.

Performance: system show performance, net show stats, and disk show stats reveal bottlenecks; check CPU, memory, and disk utilization together.

Logs: review /ddvar/log/messages, ddfs.info, and the autosupport bundle; generate a support bundle with support bundle generate for escalation.

Common issues: replication lag (check link bandwidth and throttle), space full (run filesys clean and review retention), and authentication failures (verify NTP and AD/Kerberos).

Alerts: configure alert notifications with alerts add and verify with alerts show; SNMP traps can be sent to a monitoring platform.

Verification of backups: use ddboost show stats and the backup application's verification; confirm dedup ratio and throughput trends over time.

Upgrades: check the DD OS release notes for supported upgrade paths; use system upgrade and verify with system show version before and after.

Troubleshooting methodology: isolate layer (physical link, IP, service port, application), confirm with ping/traceroute and port checks, then review logs and escalate with a support bundle.

Data Domain systems run the DD OS on dedicated appliance hardware; the Data Management Engine (DME) handles file system metadata while the Data Movement Engine streams data to disk.

Deduplication is inline and variable-length: the system segments the stream, hashes segments, and stores only unique segments in containers; the index (segment fingerprint lookup) is kept in memory and on disk.

Compression runs after deduplication; local compression plus optional global compression reduce the unique-segment footprint, so effective capacity = logical data / (dedup ratio x compression ratio).

The Data Invulnerability Architecture includes checksums on every block, write verification, and continuous fault detection with background verification (scrubbing) of stored data.

DD3300, DD6300, DD6800, DD9300, DD9500, DD9800 and the DD6900/DD9400/DD9900 generations differ in CPU, memory, and maximum shelf count; capacity scales by adding external ES/DS shelves.

Cache tiers: NVRAM holds the write cache and metadata journal for power-loss protection; SSDs (or cache devices) accelerate metadata and index lookups on larger models.

Front-end connectivity options include 10/25/40/100 GbE Ethernet and Fibre Channel HBAs; back-end SAS or NVMe paths connect controller to disk shelves.

The DD System Manager GUI, the DD CLI (SSH), and REST API are the three administrative interfaces; the CLI is the authoritative path for most configuration and troubleshooting tasks.

High availability pairs (DD9800/DD9900-class) use active-passive controllers with failover; single-node systems rely on replication and backup copies for availability.

Sizing inputs: initial logical data volume, expected daily change rate, retention period, required dedup ratio, and growth forecast; usable capacity must cover full backups plus retention of changed data.

Rule of thumb for capacity: usable capacity >= (full backup size x number of retained fulls) / expected dedup ratio, then apply a headroom margin (commonly 20-30 percent) for metadata and scrub space.

Rack planning: verify rack depth and rail kit compatibility, weight per shelf with drives populated, and floor loading; a fully populated high-end shelf can exceed 60 kg.

Power: provide redundant PDUs on separate feeds, confirm voltage/frequency (200-240 V for most high-end models), and calculate circuit load from the model's published power draw at expected load.

Cooling and airflow: front-to-back airflow requires cold aisle at the front; maintain clearance and ambient temperature within the published operating range (typically 10-35 C).

Network planning: reserve separate VLANs/subnets for management, backup data, and replication traffic; replication links need bandwidth >= change rate per replication window.

Cabling: document Ethernet/FC port mapping, use supported optics and cable lengths, and label both ends; back-end shelf cabling must follow the model's supported topology.

Site readiness checklist: DNS/NTP reachability, IP addresses and gateways, firewall ports (for example 22, 443, 2049, 2052, 111, 3009 for replication), and licensed feature keys on hand.

First boot uses the serial console or the initial GUI wizard to set the system hostname, IP address, netmask, gateway, DNS servers, and admin password.

The sysadmin account is the default full-privilege account; create role-based accounts (admin, user, backup-operator, security-officer) rather than sharing sysadmin.

Set the system time source with NTP (ntp add server then enable ntp); Kerberos, AD, and replication depend on clock accuracy.

License installation: apply the license file with license add or via GUI; verify with license show; features include replication, encryption, extended retention, and DD Boost.

Configure the system identity and location with system set and set the admin email/alerting so SNMP traps and autosupport emails reach operations.

Enable autosupport (support autosupport notify) and configure the mail server; this drives proactive case creation and weekly reporting.

Run config setup (interactive wizard) or configure settings individually; verify with config show and net show, then commit with net apply.

Security baseline: change default passwords, disable unused services, enable FIPS mode if required, and configure the firewall (adminaccess) to restrict management subnets.

Interfaces are configured with net config [] and activated with net apply; verify with net show configuration and net show stats.

Bonding/link aggregation: use LACP (mode 4) for switch-supported aggregation; balance-alb (mode 6) works without switch configuration but is less common in production.

VLAN tagging: create a VLAN interface with net create interface vlan. and attach to the physical parent; multiple VLANs can share one physical link.

Replication traffic should use dedicated interfaces; configure replication with replication add and set the replication port (default 2051 for DD Replicator control, data on 2052).

Routing: add static routes with net route add / ; the default gateway is set per interface or via net route add default.

DNS: net set dns ; verify resolution with net lookup ; DNS failure breaks CIFS/AD joins and replication by hostname.

Firewall: adminaccess add restricts management access; verify allowed subnets with adminaccess show to avoid lockout.

Troubleshooting tools: net ping, net traceroute, net lookup, and net show stats for errors/drops; check interface status with net show hardware.

Storage is organized into the active tier (dedup storage on disk), optional archive tier, and cloud tier (for supported models), enabling tiering of cold data to object storage.

External shelves are added with storage add enclosure ; verify with storage show all and check that the shelf is licensed and detected.

Filesystem capacity is reported by filesys show space; the key figure is the percentage of usable capacity consumed, and the system enters read-only mode at 100 percent (or when the threshold is reached).

Licensing model: capacity-based licenses (per TB of usable or raw capacity) plus feature licenses; verify installed licenses with license show and check expiration.

Dedup ratio = logical (pre-dedup) data / physical (post-dedup) data; global compression adds to this, so the effective ratio reported by filesys show compression combines both.

Cleanup and garbage collection: filesys clean runs reclamation of unreferenced segments; schedule during low I/O; space is not returned until cleanup runs.

Capacity alarms: configure alerts at thresholds (for example 70/85/95 percent) via alerts add; monitor with filesys show space and system show performance.

Cloud tier requires a validated object store and a cloud unit license; movement is policy-driven (data movement policy) and data is rehydrated on recall.

DD Boost distributes part of the dedup work to the backup server (client-side dedup), reducing network traffic; it uses port 2049 (and 2051/2052 for replication) and requires a DD Boost license.

DD Boost supports Fibre Channel (DD Boost over FC) and IP; configure storage units (ddboost storage-unit create) to isolate clients and set quotas.

NFS exports are created with nfs add / [options]; verify with nfs show; NFSv3 and NFSv4 are supported depending on DD OS version.

CIFS/SMB shares are created with cifs share create path and require the system to join AD with cifs setup or authentication commands.

VTL (virtual tape library) presents emulated tape libraries and drives over FC; create with vtl add library and map drives to initiators.

NDMP supports tape-less backup to DD and three-way NDMP; enable with ndmp and configure users and port 10000.

Client integration: install the DD Boost plug-in or use native protocols (NFS/CIFS) with the backup application; verify connectivity with ddboost show connections and status.

Access control: define clients with ddboost clients add and set per-client authentication; restrict NFS exports to specific subnets and avoid no_root_squash unless required.

DD Replicator is a licensed feature; configure with replication add and verify with replication show config.

Replication can be directory-level (MTree), collection-based, or full-system; MTree replication replicates only the specified directory tree and is the most common for backup data.

Replication pairs are defined source-to-destination; initialize with replication initialize and monitor with replication show status and replication show stats.

Bandwidth throttling: set with replication throttle and schedule windows so replication does not saturate backup links; recovery point objective drives the required bandwidth.

RPO calculation: bandwidth >= (daily change rate) / (replication window in seconds); the initial baseline copy transfers full data and may need a longer window.

Replication topology: one-to-one, one-to-many (fan-out), many-to-one, and cascaded; choose based on DR site count and link capacity.

Failover: promote the destination by breaking the replication context (replication break) and reconfiguring clients; plan the reverse replication after recovery.

Security: replication traffic can be encrypted with replication encryption; verify with replication show config and confirm ports 2051/2052 are open between sites.

Health checks: system show hardware, system show alerts, filesys show space, and alerts show current give a quick status snapshot.

Performance: system show performance, net show stats, and disk show stats reveal bottlenecks; check CPU, memory, and disk utilization together.

Logs: review /ddvar/log/messages, ddfs.info, and the autosupport bundle; generate a support bundle with support bundle generate for escalation.

Common issues: replication lag (check link bandwidth and throttle), space full (run filesys clean and review retention), and authentication failures (verify NTP and AD/Kerberos).

Alerts: configure alert notifications with alerts add and verify with alerts show; SNMP traps can be sent to a monitoring platform.

Verification of backups: use ddboost show stats and the backup application's verification; confirm dedup ratio and throughput trends over time.

Upgrades: check the DD OS release notes for supported upgrade paths; use system upgrade and verify with system show version before and after.

Troubleshooting methodology: isolate layer (physical link, IP, service port, application), confirm with ping/traceroute and port checks, then review logs and escalate with a support bundle.

Data Domain systems run the DD OS on dedicated appliance hardware; the Data Management Engine (DME) handles file system metadata while the Data Movement Engine streams data to disk.

Deduplication is inline and variable-length: the system segments the stream, hashes segments, and stores only unique segments in containers; the index (segment fingerprint lookup) is kept in memory and on disk.

Compression runs after deduplication; local compression plus optional global compression reduce the unique-segment footprint, so effective capacity = logical data / (dedup ratio x compression ratio).

The Data Invulnerability Architecture includes checksums on every block, write verification, and continuous fault detection with background verification (scrubbing) of stored data.

DD3300, DD6300, DD6800, DD9300, DD9500, DD9800 and the DD6900/DD9400/DD9900 generations differ in CPU, memory, and maximum shelf count; capacity scales by adding external ES/DS shelves.

Cache tiers: NVRAM holds the write cache and metadata journal for power-loss protection; SSDs (or cache devices) accelerate metadata and index lookups on larger models.

Front-end connectivity options include 10/25/40/100 GbE Ethernet and Fibre Channel HBAs; back-end SAS or NVMe paths connect controller to disk shelves.

The DD System Manager GUI, the DD CLI (SSH), and REST API are the three administrative interfaces; the CLI is the authoritative path for most configuration and troubleshooting tasks.

High availability pairs (DD9800/DD9900-class) use active-passive controllers with failover; single-node systems rely on replication and backup copies for availability.

Sizing inputs: initial logical data volume, expected daily change rate, retention period, required dedup ratio, and growth forecast; usable capacity must cover full backups plus retention of changed data.

Rule of thumb for capacity: usable capacity >= (full backup size x number of retained fulls) / expected dedup ratio, then apply a headroom margin (commonly 20-30 percent) for metadata and scrub space.

Rack planning: verify rack depth and rail kit compatibility, weight per shelf with drives populated, and floor loading; a fully populated high-end shelf can exceed 60 kg.

Power: provide redundant PDUs on separate feeds, confirm voltage/frequency (200-240 V for most high-end models), and calculate circuit load from the model's published power draw at expected load.

Cooling and airflow: front-to-back airflow requires cold aisle at the front; maintain clearance and ambient temperature within the published operating range (typically 10-35 C).

Network planning: reserve separate VLANs/subnets for management, backup data, and replication traffic; replication links need bandwidth >= change rate per replication window.

Cabling: document Ethernet/FC port mapping, use supported optics and cable lengths, and label both ends; back-end shelf cabling must follow the model's supported topology.

Site readiness checklist: DNS/NTP reachability, IP addresses and gateways, firewall ports (for example 22, 443, 2049, 2052, 111, 3009 for replication), and licensed feature keys on hand.

First boot uses the serial console or the initial GUI wizard to set the system hostname, IP address, netmask, gateway, DNS servers, and admin password.

The sysadmin account is the default full-privilege account; create role-based accounts (admin, user, backup-operator, security-officer) rather than sharing sysadmin.

Set the system time source with NTP (ntp add server then enable ntp); Kerberos, AD, and replication depend on clock accuracy.

License installation: apply the license file with license add or via GUI; verify with license show; features include replication, encryption, extended retention, and DD Boost.

Configure the system identity and location with system set and set the admin email/alerting so SNMP traps and autosupport emails reach operations.

Enable autosupport (support autosupport notify) and configure the mail server; this drives proactive case creation and weekly reporting.

Run config setup (interactive wizard) or configure settings individually; verify with config show and net show, then commit with net apply.

Security baseline: change default passwords, disable unused services, enable FIPS mode if required, and configure the firewall (adminaccess) to restrict management subnets.

Interfaces are configured with net config [] and activated with net apply; verify with net show configuration and net show stats.

Bonding/link aggregation: use LACP (mode 4) for switch-supported aggregation; balance-alb (mode 6) works without switch configuration but is less common in production.

VLAN tagging: create a VLAN interface with net create interface vlan. and attach to the physical parent; multiple VLANs can share one physical link.

Replication traffic should use dedicated interfaces; configure replication with replication add and set the replication port (default 2051 for DD Replicator control, data on 2052).

Routing: add static routes with net route add / ; the default gateway is set per interface or via net route add default.

DNS: net set dns ; verify resolution with net lookup ; DNS failure breaks CIFS/AD joins and replication by hostname.

Firewall: adminaccess add restricts management access; verify allowed subnets with adminaccess show to avoid lockout.

Troubleshooting tools: net ping, net traceroute, net lookup, and net show stats for errors/drops; check interface status with net show hardware.

Storage is organized into the active tier (dedup storage on disk), optional archive tier, and cloud tier (for supported models), enabling tiering of cold data to object storage.

External shelves are added with storage add enclosure ; verify with storage show all and check that the shelf is licensed and detected.

Filesystem capacity is reported by filesys show space; the key figure is the percentage of usable capacity consumed, and the system enters read-only mode at 100 percent (or when the threshold is reached).

Licensing model: capacity-based licenses (per TB of usable or raw capacity) plus feature licenses; verify installed licenses with license show and check expiration.

Dedup ratio = logical (pre-dedup) data / physical (post-dedup) data; global compression adds to this, so the effective ratio reported by filesys show compression combines both.

Cleanup and garbage collection: filesys clean runs reclamation of unreferenced segments; schedule during low I/O; space is not returned until cleanup runs.

Capacity alarms: configure alerts at thresholds (for example 70/85/95 percent) via alerts add; monitor with filesys show space and system show performance.

Cloud tier requires a validated object store and a cloud unit license; movement is policy-driven (data movement policy) and data is rehydrated on recall.

DD Boost distributes part of the dedup work to the backup server (client-side dedup), reducing network traffic; it uses port 2049 (and 2051/2052 for replication) and requires a DD Boost license.

DD Boost supports Fibre Channel (DD Boost over FC) and IP; configure storage units (ddboost storage-unit create) to isolate clients and set quotas.

NFS exports are created with nfs add / [options]; verify with nfs show; NFSv3 and NFSv4 are supported depending on DD OS version.

CIFS/SMB shares are created with cifs share create path and require the system to join AD with cifs setup or authentication commands.

VTL (virtual tape library) presents emulated tape libraries and drives over FC; create with vtl add library and map drives to initiators.

NDMP supports tape-less backup to DD and three-way NDMP; enable with ndmp and configure users and port 10000.

Client integration: install the DD Boost plug-in or use native protocols (NFS/CIFS) with the backup application; verify connectivity with ddboost show connections and status.

Access control: define clients with ddboost clients add and set per-client authentication; restrict NFS exports to specific subnets and avoid no_root_squash unless required.

DD Replicator is a licensed feature; configure with replication add and verify with replication show config.

Replication can be directory-level (MTree), collection-based, or full-system; MTree replication replicates only the specified directory tree and is the most common for backup data.

Replication pairs are defined source-to-destination; initialize with replication initialize and monitor with replication show status and replication show stats.

Bandwidth throttling: set with replication throttle and schedule windows so replication does not saturate backup links; recovery point objective drives the required bandwidth.

RPO calculation: bandwidth >= (daily change rate) / (replication window in seconds); the initial baseline copy transfers full data and may need a longer window.

Replication topology: one-to-one, one-to-many (fan-out), many-to-one, and cascaded; choose based on DR site count and link capacity.

Failover: promote the destination by breaking the replication context (replication break) and reconfiguring clients; plan the reverse replication after recovery.

Security: replication traffic can be encrypted with replication encryption; verify with replication show config and confirm ports 2051/2052 are open between sites.

Health checks: system show hardware, system show alerts, filesys show space, and alerts show current give a quick status snapshot.

Performance: system show performance, net show stats, and disk show stats reveal bottlenecks; check CPU, memory, and disk utilization together.

Logs: review /ddvar/log/messages, ddfs.info, and the autosupport bundle; generate a support bundle with support bundle generate for escalation.

Common issues: replication lag (check link bandwidth and throttle), space full (run filesys clean and review retention), and authentication failures (verify NTP and AD/Kerberos).

Alerts: configure alert notifications with alerts add and verify with alerts show; SNMP traps can be sent to a monitoring platform.

Verification of backups: use ddboost show stats and the backup application's verification; confirm dedup ratio and throughput trends over time.

Upgrades: check the DD OS release notes for supported upgrade paths; use system upgrade and verify with system show version before and after.

Troubleshooting methodology: isolate layer (physical link, IP, service port, application), confirm with ping/traceroute and port checks, then review logs and escalate with a support bundle.