1 11/06/86  disk_meters
  2 
  3 Syntax as a command:  disk_meters {subsystems} {-control_args}
  4 
  5 
  6 Function:  Prints metering output from Multics disk management.
  7 
  8 
  9 Arguments:
 10 subsystems
 11    Multiple position-independant subsystem identifiers may be
 12    specified to select specific subsystem information.  If no
 13    subsystem identifiers are supplied, all subsystems are
 14    listed.  A subsystem identifier can be a starname and takes
 15    the form of:
 16 
 17       disk_meters dska
 18 
 19    Subsystem identifiers can also be supplied with the -subsystem
 20    control argument.
 21 
 22 
 23 Control arguments:
 24 -channels, -chn
 25       Requests subsystem channel information.  This will be of the
 26    form:
 27 
 28       Channel Information
 29                   Term by  Interrupt   get_io  Term w/o
 30         Connects    RUN    w/o term   w/o term  Active   Status
 31      B36  229327       26         26
 32      A38   35334        7          6
 33      B37    3729        1          1
 34      A39     132
 35 
 36 
 37    Connects         - Number of channel connections made.
 38    Interrupt w/o term - Number of interrupts without terminate
 39                       status on the interrupting channel.
 40    Term by RUN      - Number of IO's seen by RUN polling, this
 41                       is a non-interrupt drive service.
 42    get_io w/o term  - Number of io_manager calls not returning
 43                       terminate.
 44    Term w/o Active  - Number of interrupts with terminate on an
 45                       inactive channel.
 46 
 47    Status           - Current status of the channel can be one
 48                       of:
 49 
 50       IOI           - The channel has been released to IOI use.
 51       INOP          - The channel is deemed to be inoperative.
 52       BROKEN        - The channel is deemed to be broken.
 53 
 54 
 55 -detail, -dtl
 56    Requests detailed printout of drive information.  It is of
 57    the form:
 58 
 59   dska_16: PV-rpv             of LV-root
 60           #Seeks AveSeek Queue-wait Channel-wait Queued Multiplier
 61   PageRd  124498  135.53       57.0  9.5%   37.6      0       79.9
 62   PageWt   37195  145.42      140.3  2.9%   38.6      0    49754.8
 63   VtocRd   10578  159.86       37.7  0.7%   31.4      0       23.8
 64   VtocWt    6036  103.15       45.6  0.3%   24.5      0       54.5
 65   TEST     0 UNLOADs,  53 TESTs
 66  Channels 13.39% busy,  27 Combs,     3.6 IO/second.
 67 
 68 
 69 -device, -dv
 70    Followed by a list of device starnames causes selection of
 71    those devices for display.
 72 
 73 -drive, -drv
 74    Selects drive information to be printed.  This is the normal
 75    default unless -system, -channel or -header is selected. (See
 76    Examples)
 77 
 78 -header, -he
 79    Selects header information to be printed, containing lock and
 80    error information.  This is the normal default unless
 81    -system, -channel or -drive is selected. (See Examples).
 82 
 83 
 84 -logical_volume, -lv
 85    Followed by a list of logical volume starnames causes
 86    selection of drives which are members of those logical
 87    volumes for display.
 88 
 89 -long, -lg
 90    Requests all of -dtl, -chn, -q, -sys, -drv, -he.
 91 
 92 
 93 -physical_volume, -pv
 94    Followed by a list of physical volume starnames causes
 95    selection of those physical volumes for display.
 96 
 97 -queue, -q
 98    Requests inclusion of drive queue information, of the form:
 99 
100  dska_04 Queue: Ave 16.1, Alloc 99, Max Depth 50/280, Cur Depth 0
101 
102    This indicates the average queue depth for the specified
103    number of queue allocations, the maximum depth since
104    max_depth_meters were last reset and the current depth in the
105    queue.  Requests are only queued if a drive is busy and/or it
106    already has requests queued.
107 
108 
109 -report_reset, -rr
110    Requests normal statistics to be printed, according to the
111    other control arguments, and then meters to be reset to this
112    point in time (see reset).
113 
114 -reset, -rs
115    Requests that disk_meters reset its meters to this point in
116    time, and not print statistics.  A reset is accomplished by
117    making a copy of the statistics as of the reset time; future
118    invocations of the command will display the difference
119    between current statistics and the copy.
120 
121 -subsystem, -sub
122    Followed by a list of subsystem starnames causes all the drives
123    which are members of those subsystems to be selected for
124    display.
125 
126 
127 -system, -sys
128    Requests that system statistics and optimizing information be
129    printed, in the form:
130 
131  FREE Queue: Ave 9.9, Alloc 31060, Max Depth 53/272, Cur Depth 15
132       Stagnate time 5.000 seconds, 32228 PL/1 interrupt services.
133       Maximum Depth Meters reset at: 11/26/84  2026.9 mst Mon
134     PageRd   Max Load    6, Depth   5 (PageRd), Fraction 0.1667
135     PageWt   Max Load 2040, Depth   2 (PageWt), Fraction 0.9990
136     VtocRd   Max Load    6, Depth   0 (VtocRd), Fraction 1.0000
137     VtocWt   Max Load   12, Depth   0 (VtocWt), Fraction 1.0000
138     BootRd   Max Load    6, Depth   0 (BootRd), Fraction 0.0000
139     BootWt   Max Load   12, Depth   0 (BootWt), Fraction 0.0000
140 
141 
142    This indicates FREE Queue use, stagnation time beyond which
143    the system does disk combing and the number of times that the
144    ALM driver had to call the PL1 driver to process complex
145    interrupt information.  The time that max_depth meters were
146    last reset at is given, as is the current status of the
147    system-wide load optimization algorithm.
148 
149 -unreset, -urs
150    Requests  that  disk_meters  reset its meters to boot time, by
151    releasing its temporary meters segment.
152 
153 
154 Examples:
155    The default invocation of disk_meters will provide information
156    using -header and -drive. The first set of information displayed
157    is the -header information for each selected subsystem. This is
158    followed by the -drive information.
159 
160    The -header information is displayed for each subsystem as follows:
161 
162  Subsystem  dska: 4 Errors 1 EDAC Error
163                   Locks   Waits    %Calls  Average      %CPU
164   Call Lock:     969088   20686   2.1346%    0.610   0.02466%
165   Run Lock:        5946     101   1.6986%    0.399   0.00008%
166   Int Lock:      998154   20608   2.0646%    0.796   0.03207%
167   Alloc Lock:    968813       0   0.0000%    0.000   0.00000%
168 
169    The -drive information is displayed for each device of the selected
170    subsystems. The display varies dependent on device type.
171 
172    For MSU451 devices the display will be like:
173 
174  Drive      PV                          Ave     ATB     ATB        ATB
175                        Reads   Writes  Seek    Reads   Writes      I/O
176 dska_01  pub01          4194     2762   160      671     1019      404
177 dska_02  list01           34             38    82854             82854
178 dska_03  pub07          3724     2309   169      756     1220      466
179 dska_04  pub04          3565     2358   160      790     1194      475
180 
181   This indicates the metering period, the subsystem and lock
182 information for the subsystem, and individual drive IO information for
183 all drives which have performed IO in the metering period.  Typically 0
184 counts are suppressed to highlight useful information.  The drive name
185 and physical volume name are included in the output.
186 
187    For MSU500, 501 devices the display will be like:
188 
189  Drive      PV                          Ave     ATB     ATB        ATB
190                        Reads   Writes  Seek    Reads   Writes      I/O
191 dskb_11  listings01     1949     1650    83     1445     1707      406
192 dskb_12  listings02     1488     1842    96     1893     1529
193 dskb_13  alpha01        2066     1796    28     1363     1568      404
194 dskb_14  alpha02        1671     1437    31     1685     1960
195 dskb_17  pdir03         1912     1988    20     1473     1417      366
196 dskb_18  pdir04         1939     1850    23     1452     1522
197 
198   In this example, meters are displayed for primary and secondary
199 devices, the ATB I/O is the ATB I/O of the physical spindle, rather
200 than of the logical device.  The ATB for read and write are the ATB on
201 the logical device.  ATB I/O appears only on the line for the primary
202 device of the spindle.
203 
204   For MSU3380, 3381 devices:
205 
206  Drive      PV                          Ave     ATB     ATB        ATB
207                        Reads   Writes  Seek    Reads   Writes      I/O
208 dskd_00a fpdir01
209 dskd_00b fpdir02      10653     8543    73      264      329      146
210 dskd_01a fpdir03
211 dskd_01b fpdir04       4100     4051    34      687      695      345
212 dskc_02a root2
213 dskc_02b
214 dskc_02c                277      372    62     4241     3158     1810
215 dskc_03a root3
216 dskc_03b
217 dskc_03c                223      416    60     5268     2824     1838
218 
219  In this example meters are kept and displayed for the entire device.
220 However, each subvolume name of the device is displayed with the
221 physical volume name, if one is assigned.  The metering information
222 appears on the line for the last subvolume.
223