HP StorageWorks Command View SDM Installation and User Guide (T1086-96023, February 2005)

168 Command Line User Interface
Protocol Error Count FRONTEND-
PORT,
BACKEND-
PORT
The number of N_Port protocol errors detected and LRR
primitive sequences received while the link was up.
Usage
Each of the port error counters provides a general
indication of the quality of port operation. Any significant
counts in any of the port error counters is an indication
that port or link operational quality may be having an
effect on performance.
Compatibility
FRONTEND-PORT metrics are available for all products.
BACKEND-PORT metrics are available for VA 7410 only.
Queue Full Statuses Sent (1.05
release)
Queue Full Response (prior to
1.05)
ARRAY
(7100/
7400/7110)
OPAQUE
FRONTEND-
PORT (7410)
The number of commands that were failed with a SCSI
status of “Queue Full.”
Usage
The Queue Full status indicates that the array is completely
full of work and cannot accept any more commands. A
large number of Queue Full statuses is an indication that
the array is not configured to effectively manage the host
workload. In this case users may be experiencing
unacceptably long response time when attempting to
access data on the array. If possible, adding more disks to
the array may help to improve response time.
RAID 1+0 Allocation (absolute) LUN The number of logical blocks in the LUN that are stored
using the RAID 1+0 storage method.
Usage
RAID 1+0 provides better random write performance than
RAID 5 DP. The more data stored in RAID 1+0 as
compared to RAID 5 DP, the better the random write
performance that can be achieved. This metric combined
with the RAID 5 DP Allocation metric for a LUN will
indicate the class of random write performance that can
be expected from the LUN. Some LUN's in an AutoRAID
configuration may be almost completely stored as RAID
1+0 if the data stored in the LUN is mostly within the write
working set.
Table 12 Performance Metrics (Continued)
Metric Categories Description/Usage