The first section shows the configuration of the selected storage subsystem. This includes the amount of cache configured (installed) and available for use, and how much storage, if any, is pinned because of a DASD failure. It also includes the amount of non-volatile storage needed for the DASD FAST WRITE function. Finally, the overall subsystem status is shown in terms of whether caching is active, or, for example, CACHE FAST WRITE is activated.
The second section provides details of the cache usage of the subsystem. Here, the analyst can see at a glance all the I/O requests to the subsystem, divided into the categories NORMAL, SEQUENTIAL, and CACHE FAST WRITE. READ and WRITE requests are shown separately as totals, rates, or ratios.
C A C H E S U B S Y S T E M A C T I V I T Y
PAGE 1
z/OS V2R1 SYSTEM ID SYS1 DATE 09/28/2013 INTERVAL 15.00.000
RPT VERSION V2R1 RMF TIME 00.30.00
SUBSYSTEM 2105-01 CU-ID 1E91 SSID 3007 CDATE 09/28/2013 CTIME 00.29.55 CINT 15.00
TYPE-MODEL 2105-E20 MANUF IBM PLANT 75 SERIAL 000000016374
------------------------------------------------------------------------------------------------------------------------------------
CACHE SUBSYSTEM STATUS
------------------------------------------------------------------------------------------------------------------------------------
SUBSYSTEM STORAGE NON-VOLATILE STORAGE STATUS
CONFIGURED 3072.0M CONFIGURED 192.0M CACHING - ACTIVE
AVAILABLE 2692.2M PINNED 0.0 NON-VOLATILE STORAGE - ACTIVE
PINNED 0.0 CACHE FAST WRITE - ACTIVE
OFFLINE 0.0 IML DEVICE AVAILABLE - YES
------------------------------------------------------------------------------------------------------------------------------------
CACHE SUBSYSTEM OVERVIEW
------------------------------------------------------------------------------------------------------------------------------------
TOTAL I/O 185787 CACHE I/O 185787 CACHE OFFLINE 0
TOTAL H/R 0.893 CACHE H/R 0.893
CACHE I/O -------------READ I/O REQUESTS------------- ----------------------WRITE I/O REQUESTS---------------------- %
REQUESTS COUNT RATE HITS RATE H/R COUNT RATE FAST RATE HITS RATE H/R READ
NORMAL 151012 167.0 132096 146.1 0.875 5339 5.9 5339 5.9 5339 5.9 1.000 96.6
SEQUENTIAL 24588 27.2 23696 26.2 0.964 4848 5.4 4848 5.4 4848 5.4 1.000 83.5
CFW DATA 0 0.0 0 0.0 N/A 0 0.0 0 0.0 0 0.0 N/A N/A
TOTAL 175600 194.2 155792 172.3 0.887 10187 11.3 10187 11.3 10187 11.3 1.000 94.5
-----------------------CACHE MISSES----------------------- ------------MISC------------ ------NON-CACHE I/O-----
REQUESTS READ RATE WRITE RATE TRACKS RATE COUNT RATE COUNT RATE
DFW BYPASS 0 0.0 ICL 0 0.0
NORMAL 18916 20.9 0 0.0 25086 27.8 CFW BYPASS 0 0.0 BYPASS 0 0.0
SEQUENTIAL 892 1.0 0 0.0 36075 39.9 DFW INHIBIT 0 0.0 TOTAL 0 0.0
CFW DATA 0 0.0 0 0.0 ASYNC (TRKS) 18546 20.5
TOTAL 19808 RATE 21.9
---CKD STATISTICS--- ---RECORD CACHING--- ----HOST ADAPTER ACTIVITY--- --------DISK ACTIVITY-------
BYTES BYTES RESP BYTES BYTES
WRITE 0 READ MISSES 0 /REQ /SEC TIME /REQ /SEC
WRITE HITS 0 WRITE PROM 472 READ 982 1.1K READ 16.912 65.9K 12.7M
WRITE 15.9K 2.8M WRITE 12.465 48.3K 4.1M
| Field Heading | Meaning |
|---|---|
| SUBSYSTEM | Storage subsystem type (as configured). |
| CU-ID | Physical control unit number of the caching subsystem. This is equal the lowest device number, or to the device that has been turned online first, respectively. |
| SSID | Subsystem identifier: a number assigned during installation of the subsystem that uniquely identifies the storage subsystem. |
| Note: Device reserve activity
can cause a data gatherer interface to wait until a reserve has been
released. This in turn can cause the cache interval to be much longer
than a regular RMF interval.
Therefore, CDATE, CTIME and CINT have been introduced to show the actual point in time to which the cache interval start is related, and the actual cache interval length. All rates shown in the report are based on CINT, not on INTERVAL. |
|
| CDATE | Date when the cache interval started. |
| CTIME | Time when the cache interval started. |
| CINT | Cache interval time. In interval reports, the format is mm.ss, while in duration reports the format is hh.mm.ss. |
| TYPE-MODEL | Device type and model. |
| TYPE MODEL MANUF PLANT SERIAL | The hardware description of the disk system. |
| VOLSER | Volume serial number of the reported DASD device (only for device-level reporting). |
| NUM | Device number of the reported DASD device (only for device-level reporting). |
| Field Heading | Meaning |
|---|---|
| Note: The reported storage capacities
for cache and non-volatile storage (NVS) represents only the Cluster
Processor Complex in the storage server, that controls the subsystem.
Since a typical storage server has two clusters, you must double the
reported capacities to get the actual sizes.
|
|
| SUBSYSTEM STORAGE | Physical capacity of random access cache (in megabytes). |
| CONFIGURED | Amount of storage that is installed in this storage subsystem. |
| AVAILABLE | Amount of storage that is available for caching. This is the total cache size minus the amount used by the subsystem for the cache directory, minus the amount pinned and offline storage. |
| PINNED | Amount of storage that is unavailable because a DASD failure is preventing the subsystem from destaging the data. The data is pinned in cache. |
| OFFLINE | Amount of storage that is offline because of a host or subsystem error. |
| NON-VOLATILE STORAGE (NVS) | Physical capacity of random access storage with a backup battery power source (in megabytes). |
| CONFIGURED | Amount of NVS that is installed in this storage subsystem. |
| PINNED | Amount of NVS that is unavailable because a DASD failure is preventing the subsystem from destaging the data. The data is pinned in NVS. |
| STATUS | The caching status of the entire subsystem. |
| CACHING | Overall caching status of the subsystem.
|
| NON-VOLATILE STORAGE | Overall status of the non-volatile storage (NVS).
|
| CACHE FAST WRITE | Status of the CACHE FAST WRITE (CFW) option.
Note: CFW does not use NVS.
|
| IML DEVICE AVAILABLE | Status of the IML device.
|
| Field Heading | Meaning |
|---|---|
| TOTAL I/O | Total number of I/O requests to cached devices in the storage subsystem. |
| TOTAL H/R | Ratio of I/Os that were processed within the cache (cache hits) based on the total number of I/Os. |
| CACHE I/O | Total number of cacheable I/O requests to cached devices in
the storage subsystem. This value excludes INHIBIT CACHE LOAD and CACHE BYPASS I/O requests. |
| CACHE H/R | Ratio of I/Os that were processed within the cache (cache hits) based on the total number of cacheable I/O requests. |
| CACHE OFFLINE | Total number of I/O requests to non-cached devices in the storage subsystem. |
| CACHE I/O REQUEST - The channel command DEFINE EXTENT specifies the way the cache will be used. There are three categories (NORMAL, SEQUENTIAL, CFW DATA) and a TOTAL value: | |
| NORMAL | Cache will be managed by least-recently-used (LRU) algorithm for making cache space available. |
| SEQUENTIAL | Tracks following the track assigned in the current CCW chain are promoted. They will be transferred from DASD to cache in anticipation of a short-term requirement. |
| CFW DATA | WRITE and READ-AFTER-WRITE requests are processed in cache. The data might not be written to DASD. Because CFW does not use the NVS, the application is responsible for restoring the data after a cache or subsystem failure. |
| TOTAL | This is either the sum of I/O requests, the total I/O rate, or the average hit ratio for the three categories described above. |
| READ I/O REQUESTS - Cache I/O requests that searched or read data from DASD. This is the number of channel operations that had at least one SEARCH or READ command but no WRITE commands. This is counted for cached devices only. | |
| COUNT | Total number of SEARCH/READ requests. |
| HITS | Number of SEARCH/READ requests that completed without accessing the DASD. |
| H/R | Hit Ratio - number of SEARCH/READ hits compared to the total number of SEARCH/READ requests. |
| WRITE I/O REQUESTS - Cache I/O requests that wrote data to DASD. This is the number of channel commands that had at least one WRITE command. It is counted for cached devices only. | |
| COUNT | Total number of WRITE requests. |
| FAST | Total number of DASD/CACHE FAST WRITE requests. |
| HITS | Number of DASD/CACHE FAST WRITE requests that completed without accessing the DASD (fast write hit). |
| H/R | Number of DASD/CACHE FAST WRITE hits compared to the sum of all READ and WRITE requests (excluding ICL and BYPASS). |
| %READ | Percentage of READ requests compared to the sum of all READ and WRITE requests (excluding ICL and BYPASS). |
| CACHE MISSES -Cache misses are
calculated as the difference between total I/O requests and the number
of cache hits. They are shown for normal, sequential, and CFW requests.
A cache miss occurs if a record that is to be read from or written to a data set is not found in the cache. In either case, the track in which this record is located is searched on the DASD and transferred to cache. Subsequent requests referring to the same record can then be processed in cache only, provided that the data has not yet been moved out because cache space is needed for other data. Other reasons for cache misses are DFW bypass and DFW inhibit. In the first case, a lack of NVS caused the data to be written to DASD immediately, while in the second case, usage of DFW is inhibited (for example, the device does not allow DFW). |
|
| READ | Number of SEARCH/READ requests that needed access to DASD because the data could not be found in the cache. |
| WRITE | Number of WRITE requests that needed access to DASD because the data could not be found in the cache. |
| TRACKS | Number of tracks transferred from DASD to cache. |
| TOTAL | Total number of I/O requests that needed access to DASD because the data could not be found in the cache. |
| MISC - Miscellaneous cache activities | |
| DFW BYPASS | Number of DASD FAST WRITE requests that would have resulted
in a DFW hit; however, NVS was overutilized causing writes to be sent
directly to DASD. This value is also known as DFW RETRY. |
| CFW BYPASS | Number of operations that did not transfer a track from DASD into cache because no free segments were available. If no free segments are available in cache, there is no destaging in favor of I/O requests with the CACHE FAST WRITE attribute. The I/O goes directly to the DASD. |
| DFW INHIBIT | If DASD FAST WRITE is active, this is the number of WRITE requests that inhibited DASD FAST WRITE. If DASD FAST WRITE is inactive, this is the number of WRITE requests that directly accessed the DASD, even with DASD FAST WRITE turned on. |
| ASYNC | Number of tracks transferred from cache to DASD asynchronously
to transfer from the channel to release space in the cache and the
NVS. For a duplex pair, this is the count of transfers from the cache to the secondary device of a duplex pair. Transfers from the cache to the primary device of a duplex pair are not counted. A high number of ASYNC I/Os is an indicator for an over-committed cache or NVS. |
| NON-CACHE I/O - READ requests that switched off cache processing. | |
| ICL | Inhibit cache load. Number of I/O requests that inhibited load
of data into cache although the data was not found in the cache. Note: If
the data had been in the cache, it would have been counted as cache
hit. Therefore, this is actually the number of ICL misses.
|
| BYPASS | Number of I/O requests that explicitly bypassed the cache, irrespective of whether the data is in the cache or not. |
| TOTAL | Total number of I/O requests that bypassed the cache. |
| CKD STATISTICS - CKD (Count-Key-Data) is a format used to store data on DASD. The counts shown in this section are contained in the total write count. | |
| WRITES | Number of write I/O requests in CKD format. |
| WRITE HITS | Number of write I/O requests in CKD format that could be resolved in the cache. |
| RECORD CACHING - Record caching is done dynamically upon a decision made by DCME or the microcode. It may improve overall cache performance if caching of whole tracks would waste cache storage. The decision is based on the number of I/Os, the hit ratio, and the locality of reference of a certain entity of data. | |
| READ MISSES | Number of instances in which a record requested for READ was not found in the cache, and access to DASD was required. |
| WRITE PROM | Number of instances in which a record requested for WRITE was found in the cache, and access to DASD was not required. |
| HOST ADAPTER ACTIVITY1) - I/O activity of normal, sequential and CFW read and write requests. | |
| BYTES/REQ | The average number of transferred bytes per read and write request. |
| BYTES/SEC | The average number of transferred bytes per second for read and write requests. |
| DISK ACTIVITY1) - Transfer activity from hard disk to cache and vice versa. | |
| RESP TIME | Response time in milliseconds per read and write request. |
| BYTES/REQ | The average number of transferred bytes per read and write request. |
| BYTES/SEC | The average number of transferred bytes per second for read and write requests. |
| 1) available for the IBM TotalStorage DS family | |
Duration reports: An asterisk behind a status field or the field VOLUME SERIAL indicates that the contents of the field has changed during the duration interval.