DASDRESP question and DASD Latency
Article ID: 125040
Updated On:
Products
Issue/Introduction
We have a question about DASDRESP that we couldn't find the answer in the manuals:
We have the MVS-DEVICES scheduled every 15 minutes;
Does the data shown on DASDRESP represent the average of those 15 mintues or a snapshot every 15 minutes.
If it's a snapshot, it's for how long?
Also, does this help to measure DASD Latency?
Environment
Component: SYSVW
Resolution
The data in DASDRESP is coming from the CMB (Channel Measurement Block).
For each DASD device, the CMB contains a cumulative SSCH+RSCH count, a Sample count, Connect time, Pending time, Disconnect time, and Control-Unit Queueing time (all times are in 128-micro seconds).
Our MVS data collector obtains the CMB data roughly every 30 seconds and then takes the delta for each of these fields.
The SSCHSamp and SSCHCnt fields show the delta values
The time fields are converted to seconds, then averaged out using the interval length.
The interval is roughly 60 seconds and will usually contain two samples per interval.
So the times shown are the averages over the interval.
The DASDRESP command can be used, along with the DELAY command to monitor DASD Latency. Fields are as follows:
+------------------+-------------------------------------------------------------+
| Service | The total device service time. |
| | |
| | service = connect + disconnect + pend + queue |
+------------------+--------------------------------------------------------------+
| Connect | The device connect time. |
| | |
| | The average time the device was connected to a |
| | channel path and actually transferring data |
| | between the device and central storage. |
+------------------+--------------------------------------------------------------+
| Disconn | The subchannel disconnect time. |
| | |
| | The average time the device was disconnected |
| | (not transferring data) while processing an |
| | SSCH instruction. |
+------------------+---------------------------------------------------------------+
| Pend | The SSCH request pending time. |
| | |
| | The average time an I/O request remains queued |
| | in the channel. |
+------------------+---------------------------------------------------------------+
| Queu | The control unit queue time. |
| | |
| | The average time an I/O request must wait on |
| | an IOS queue before a SSCH instruction can be |
| | issued. |
+------------------+----------------------------------------------------------------+
| Active | The device active time. |
| | |
+------------------+----------------------------------------------------------------+
| CmdResp | The device initial command response time. |
| | |
| | The average delay time that a successfully |
| | initiated start or resume function needs |
| | until the first command is indicated as |
| | being accepted by the device. |
--------------------------------------------------------------------------------------
From the screen Select a device, and put a ? on the device's command line which will show other commands that can be used to investigate that device. It may be helpful on the initial DASDRESP screen to sort on one of the fields, like the CmdResp field for example, first.
Another screen is the DELAYS screen. Do a SORT REASON D and look for any with reason DEVICE. From here Select the device to get more information.
Feedback
