Note in this minidisk cache analysis that during the stress test between 7am and 8am, minidisk cache was very large. The storage requirements and paging requirements allowed VM to utilize real storage for cache instead of user pages.
Field Descriptions: Most of the fields on the ESAMDC report should be self explanatory. The I/O per second shows the dasd I/O requests and the subset of those that were hits, satisfied by the cache. Main storage is managed by CP with targets for minimum and maximum values set by the installation - and an objective set by CP within the boundaries based on current load. Some fields that might not be intuitive are the insertion data fields. An insert into cache is performed on behalf of a user. Each user has a maximum fair share of 4K blocks in minidisk cache. Note that as more users logon to the system this value gets smaller. Insertions not allowed (NotAld) are not allowed because the user has exceeded the maximum value.
Report: ESAMDC Minidisk Cache Analysis Velocit
Monitor initialized: on 7490 serial 00001 First r
--------------------------------------------------------------------------
<----Load----> <-----Main Storage MDC-->
<-Users-> Tran Hit Usr Per Not <-Sizes (MB)-->
Time Actv In Q /sec Pct rds hits Max Min Ald Avg MIN MAX Obj Stls Delt
-------- ---- ---- ---- --- --- ---- --- --- --- --- --- --- --- ---- ----
03/10/01
01:00:01 9 2.2 0.4 75 12 8.6 2K 0.6 0 41 0 128 89 0.1 16.0
02:00:01 8 1.7 0.3 98 3.2 3.1 3K 0.1 0 17 0 128 92 0 0
03:00:01 9 1.8 0.4 97 5.1 4.9 3K 0.1 0 28 0 128 92 0 0
04:00:01 10 2.2 0.4 98 6.1 6.0 3K 0.1 0 49 0 128 92 0 0
05:00:01 9 2.0 0.4 97 4.5 4.4 3K 0.2 0 49 0 128 88 0.0 2.0
06:00:01 10 2.7 0.4 99 7.0 6.9 3K 0.2 0 40 0 128 92 0 0
07:00:01 11 2.8 0.4 90 6.3 5.6 2K 0.3 0 51 0 128 88 0.0 5.2
08:00:01 11 3.2 0.4 90 5.4 4.8 1K 1.1 0 75 0 128 77 0.0 10.3
**************************************************Summary*****************
Average: 10 2.3 0.4 91 6.1 5.5 3K 0.3 0 44 0 128 89 0.0 4.2