Operational definitions — the way engineers actually use these terms on a bridge call. Concise, accurate, and grouped by domain.
[VIRTUALIZATION]
Virtualization
PSOD
Purple Screen of Death — ESXi kernel panic. Stack trace must be captured before reboot or root cause is usually unrecoverable.
APD / PDL
All Paths Down / Permanent Device Loss — ESXi storage states that determine whether the host waits or fails fast on missing LUNs.
vSAN witness
Third-site quorum component for stretched vSAN clusters. Its loss alone does not break production but eliminates DR capability.
Admission control
HA policy that reserves cluster capacity for failover. Misconfiguration is the most common cause of 'HA didn't restart my VM'.
VMFS resignature
Re-presenting a VMFS datastore with a new UUID after snapshot or replication. Required after most DR failovers.
vCLS
vSphere Cluster Services — small agent VMs that keep DRS/HA functioning independently of vCenter. Do not delete.
DRS
Distributed Resource Scheduler — load-balances VMs across hosts. Aggressiveness setting decides how often it moves workload.
[STORAGE]
Storage
URE
Unrecoverable Read Error — a sector the disk cannot read. Specified per ~10^14 bits on consumer SATA; the math is why RAID 5 is risky on large arrays.
Foreign config
An array configuration the current controller did not create. Import is usually safe; initialize destroys it.
Write hole
Window during which a parity write is partially complete. Power loss here corrupts the stripe; battery-backed cache exists to prevent it.
Hot spare
A disk reserved for automatic replacement of a failed member. Useful only if health-checked regularly.
Resilver / rebuild
Rewriting a replacement disk from parity or mirror. Read-heavy on surviving disks; a common cause of cascade failures.
SHR
Synology Hybrid RAID — flexible layout allowing mixed disk sizes. Recovers like RAID 5/6 underneath.
Aggregate / FlexVol
NetApp container concepts — aggregates pool physical disks, FlexVols are flexible volumes living inside them.
[COLOCATION & NETWORK]
Colocation & Network
Smart hands
On-site engineer authorized to diagnose and make decisions, not just execute scripted physical tasks.
Remote hands
Technician executing exact, scripted instructions from your team. Faster, lower-skill, more rigid.
Cross-connect
Physical fiber or copper run between two cages or to a carrier meet-me-room. Frequently mislabeled at carrier handoff.
Crash cart
Mobile keyboard/monitor/mouse for direct console access. Facility-shared and frequently unavailable when needed.
Suite / cage / cabinet
Suite = walled private room. Cage = mesh-walled area with locked door. Cabinet = single rack with a lockable door.
MMR
Meet-Me-Room — facility-neutral space where carrier circuits hand off to customer cross-connects.
PDU
Power Distribution Unit. A and B sides should land on independent UPS feeds; if they don't, you have one PDU, not two.
[INCIDENT & OPERATIONS]
Incident & Operations
SEV-1 through SEV-4
Severity classification by blast radius and recoverability, not by alert volume. See the severity matrix on the Resources page.
RTO / RPO
Recovery Time Objective / Recovery Point Objective — how long until service is back and how much data you can afford to lose.
Blast radius
What else fails if this fault propagates. The primary input to severity classification.
CAB
Change Advisory Board — the formal approval process for non-emergency changes. SEV-1 work proceeds under emergency change authority and is documented retroactively.
Bridge
Conference call open for the duration of an incident. Engineers join, leave, and hand off here.
Chain of custody
Documented record of who touched what, when. Required for regulated environments and post-incident forensics.