SG0-001 · Question #262
A production cluster is attached to redundant disk fabrics and a tape fabric. Each cluster member has access to twelve (12) TB of RAID 5 production storage and twelve (12) TB of RAID 0 local replica a
The correct answer is A. Tape HBA. The single 2Gb/s Tape HBA is the bottleneck because the aggregated throughput of the ten LTO II tape drives (2.8 Gb/s) exceeds its capacity, even though the disk subsystem can supply data faster.
Question
A production cluster is attached to redundant disk fabrics and a tape fabric. Each cluster member has access to twelve (12) TB of RAID 5 production storage and twelve (12) TB of RAID 0 local replica accessed through eight HBAs single initiator zoned to sixteen storage ports on one storage subsystem. The tape fabric includes one connection to each server and connections to four (4) 1Gb FC to SCSI gateways connected to ten (10) LTO II (35MB/s) tape drives. The standby cluster member performs multi-threaded backups using the local replica. All HBAs and ports are 2Gb/s. If the recovery time is equal to backup time, where is the bottleneck?
Options
- ATape HBA
- BDisk HBAs
- CTape Drives
- DFC to SCSI Gateway
How the community answered
(43 responses)- A58% (25)
- B12% (5)
- C7% (3)
- D23% (10)
Why each option
The single 2Gb/s Tape HBA is the bottleneck because the aggregated throughput of the ten LTO II tape drives (2.8 Gb/s) exceeds its capacity, even though the disk subsystem can supply data faster.
The server has one 2Gb/s HBA for the tape fabric. The combined throughput of ten LTO II tape drives (35MB/s each) is 350 MB/s, which converts to 2.8 Gb/s. This 2.8 Gb/s demand exceeds the 2 Gb/s capacity of the single tape HBA, making it the bottleneck.
The server has eight 2Gb/s disk HBAs, providing a total of 16 Gb/s of bandwidth, which is significantly more than the tape subsystem's aggregate demand of 2.8 Gb/s.
While individual LTO II drives are 35MB/s, their combined capacity of 2.8 Gb/s is limited by the upstream 2Gb/s tape HBA, making the HBA the primary bottleneck before the drives themselves are fully saturated.
The four 1Gb FC to SCSI gateways provide a combined 4Gb/s of bandwidth, which is greater than the 2.8 Gb/s aggregate speed of the tape drives, indicating the bottleneck is not at the gateways but further upstream at the tape HBA.
Concept tested: SAN/backup performance bottleneck identification
Topics
Community Discussion
No community discussion yet for this question.