RDF/A Group metrics
Unisphere collects and displays metrics for SRDF/A groups.
Within the PM metrics section, V4 is used to denote that the metric is supported only by a storage system (array) running PowerMaxOS 10 (6079). V3 is used to denote that the metric is supported only by a storage system (array) running HYPERMAX OS 5977 or PowerMaxOS 5978. Metrics without the denotation are supported by all storage systems (arrays).
The following table lists all metrics that are available for SRDF/A groups.
| Metric | Description |
|---|---|
| % Compressed Throughput Received | Calculated value: ((Compressed MBs Received per sec/MBs Received per sec) * 100) |
| % Compressed Throughput Sent | Calculated value: ((Compressed MBs Sent per sec/MBs Sent per sec) * 100) |
| % Compressed Throughput Sent and Received | Calculated value: ((Compressed MBs Sent and Received per sec/MBs Sent and Received per sec) * 100) |
| Active Cycle Capacity Size | Active cycle capacity size |
| Active Cycle Size | The count of cache slots that measures the amount of data captured (on the R1 side) or received (on the R2 side) (V3) |
| Avg Cycle Size | The average size of all the cycles in SRDF/A sessions (V3) |
| Avg Cycle Size Capacity | Average cycle size capacity |
| Avg Cycle Time | The mean time (in seconds) of the last 16 cycles (calculated as true averages, for example, if only 8 cycles have occurred, the average represents those 8 cycles only) |
| Avg IO Service Time | The average time the director takes to serve I/O |
| Capacity of RDF Invalid Tracks | Capacity of SRDF invalid tracks |
| Capacity of Tracks Spilled Back | Capacity tracks spilled back |
| Capacity Tracks Spilled Over | Capacity tracks spilled over |
| Capacity Tracks Spilled Over Used | Capacity tracks spilled over used |
| Compressed Throughput Received | The number of compressed MBs/sec received by this SRDF/A group |
| Compressed Throughput Sent | The number of compressed MBs/sec sent by this SRDF/A group |
| Compressed Throughput Sent and Received | The number of compressed MBs sent and received by this SRDF/A group |
| Cycle Number | The active cycle number identifier for the given SRDF/A session - in a regular case, the cycles switch every ~30 seconds, however, usually the collection interval is in minutes. Calculating the difference in cycle numbers between collection intervals relative to the minimum cycle time shows how many cycles occurred over the last interval. |
| DSE Threshold | The percent of write pendings before DSE activates |
| DSE Threshold Capacity | DSE threshold capacity |
| DSE Used Tracks | The active cycle spilled count plus the inactive cycle spilled count (V3) |
| DSE Used Tracks Capacity | DSE used tracks capacity |
| Inactive Cycle Capacity Size | Inactive cycle capacity size |
| Inactive Cycle Size | The count of cache slots that measures the amount of data transmitted (on the R1 side) or applied (on the R2 side) (V3) |
| Local WP Count | The number of write pending slots waiting to be de-staged to disk. On the R1 - should be less than or equal to the system write pending count (V3) |
| Local WP Used Capacity | Local WP used capacity |
| Num Devs in Session | The number of storage volumes defined in this session |
| Number of Compressed Links | The number of compressed links used by this SRDF/A group |
| Number of Links | The number of links used by this SRDF/A group |
| Number of RDF Invalid Tracks | Number of SRDF invalid tracks (V3) |
| R1 to R2 Lag Time | The time that R2 is behind R1 (RPO) - this metric is calculated as the last cycle time plus the time since last switch. In a regular case, the cycles switch every ~30 sec and the samples are taken every few minutes, therefore this counter may not show significant data, however, when cycles elongate beyond the sample time, this counter can help indicate an estimate of the RPO. |
| R2 Last Restore Time | The time, in seconds, that the restore of the last active cycle took to complete. The restore is an operation that is performed by the DA and RA to mark the tracks in the apply cycle as Write Pending to the local drives. This operation is usually a quick cache operation, unless one of the volumes reaches its WP limit. The importance of this counter is that it shows one of the common reasons for an increase in the cycle time. |
| RDF R1 to R2 IOs/sec | The number of I/O operations per second that are being transmitted across the links for this session |
| RDF R1 to R2 Throughput | The MB/sec that are being transmitted across the links for this session |
| RDF R2 to R1 IOs/sec | The number of I/O operations per second that are being transmitted across the links for this session |
| RDF R2 to R1 Throughput | The MB/sec that are being transmitted across the links for this session |
| RDF Reads/sec | SRDF reads per second |
| RDF Response Time | The average time that it takes to send an I/Oover the SRDF Link - the SRDF round-trip time. The metric is calculated as follows: SRDF Accumulated Service Time /(SRDF Reads/sec + SRDF Writes/sec) rolled up from the SRDF devices in the group |
| RDF Rewrites/sec | SRDF rewrites per second |
| RDF Throughput Read | SRDF Throughput Read |
| RDF Throughput Written | SRDF Throughput Written |
| RDF Writes/sec | SRDF writes per second |
| RDF/A Active Indicator | Indicates if SRDF/A group is active (0/1) |
| RDF/A Session Indicator | Indicates if group is SRDF/A (0/1) |
| RDF/A WP Count | The number of SRDF/A writes pending |
| RDF/A WP Count Capacity | SRDF/A WP count capacity |
| Read Hits/sec | The total number of read operations that were immediately satisfied by cache. |
| Reads/sec | The number of tracks read from the R2 (host reads + copy) |
| Session Number | The number of the SRDF/A session |
| Spillover Packed Slots | Spillover packed slots (V3) |
| Spillover Packed Slots Capacity | Spill over packed slots capacity |
| Throughput Read | Throughput Read |
| Throughput Written | Throughput Written |
| Time Since Last Switch | The time (in seconds) since the last switch from active to inactive or inactive to active - since in a regular case the cycles switch every ~30 sec and the samples are taken every few minutes, this counter may not show significant data, however, when cycles elongate beyond the sample time, this counter can help indicate an estimate of the R1 to R1 lag time. |
| Total HA Repeat Writes/sec | The total number of writes to a slot already in the active cycle - total host adapter repeat writes (measured in write commands) to SRDF/A volumes only. This counter helps estimate the cache locality of reference, that is, the amount of cache that is saved by the re-writes. This does not give any indication to the bandwidth locality of reference. |
| Total HA Writes/sec | The total host adapter writes - this is measured in write commands to SRDF/A volumes only |
| Total Tracks Spilled Back | Total tracks spilled back (V3) |
| Total Tracks Spilled Over | Total tracks spilled over (V3) |
| Total Tracks Spilled Over Used | Total tracks spilled over used (V3) |
| Uncommitted Tracks | The number of tracks for all the SRDF/A volumes in the session that have not been committed to the R2 volume (V3) |
| Uncommitted Tracks Capacity | Uncommitted tracks capacity |
| Writes/sec | The number of write requests to the SRDF director per second |