![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to charge.1 CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [Research Unix] / researchv10dc / man / manx|
Return to charge.1 CVS log | Up to [Research Unix] / researchv10dc / man / manx |
1.1 root 1: .TH CHARGE 1 SHARE
2: .SH NAME
3: charge, charges \- change (show) charges for share system usage calculations
4: .SH SYNOPSIS
5: .B charge
6: [\-flags]
7: .br
8: .B charges
9: [\-flags]
10: .SH DESCRIPTION
11: .I Charge
12: will change the costs associated with the usage calculations for the
13: .IR "Share Scheduler" .
14: The flag \fB\-C\fR, if used, should be mentioned first
15: to cause
16: .I charge
17: to read the
18: existing values,
19: instead of working on default values.
20: The alternate name is used to show the charges currently in operation,
21: for which the default assumes \fB\-C\fR, unless the flag \fB\-\fP is used.
22: The optional flags are as follows:-
23: .TP "\w'-NmaxnormuXX'u"
24: .BI -
25: Causes
26: .I charges
27: to show the default settings for the charging parameters.
28: .TP
29: .BI -C
30: Causes
31: .I charge
32: to read in the current values,
33: which it will use instead of the defaults.
34: \fBMust be first flag if used\fP.
35: .TP
36: .BI -D h1,h2
37: Will set the decay rate for process priorities with normal \fInice\fP
38: so that they will decay to half their initial value in
39: .I h1
40: seconds,
41: and set the decay rate for process priorities with maximum \fInice\fP
42: so that they will decay to half their initial value in
43: .I h2
44: seconds.
45: .TP
46: .BI -E half-life
47: Will set the decay rate for users' process
48: .I rates
49: so that they will decay to half their initial value in
50: .I half-life
51: seconds.
52: .TP
53: .BI -F flags
54: Sets various global scheduling flags \(em see
55: .IR share (5)
56: for details.
57: .I Flags
58: are assumed to be in octal.
59: .TP
60: .BI -G maxgroups
61: Sets the maximum depth for the scheduling tree.
62: .TP
63: .BI -K half-life [s]
64: Will set the decay rate for users' usages
65: so that they will decay to half their initial value in
66: .I half-life
67: hours.
68: If the
69: .I half-life
70: is followed by the character \fBs\fP,
71: then the number will be interpreted as seconds.
72: .TP
73: .BI -N maxnormu
74: Upper bound on
75: .I "normalised usage"
76: used in process priority calculations.
77: The number can be added to a running process's priority every clock tick,
78: so it should be small enough not to overrun the value
79: .I maxupri
80: in too short a time interval
81: (ie: it should be less than \fI(maxupri*(1\-pri_decay)) / \s-1HZ\s0\fP,
82: see the output of \fIcharges -v\fP for the low priority value of \fIpri_decay\fP.)
83: .TP
84: .BI -P maxpri
85: Absolute upper bound for a process's priority.
86: (Something less than the largest non-negative integer.)
87: .TP
88: .BI -Q maxupri
89: Upper bound for normal processes' priorities.
90: .I Idle
91: processes run with priorities in the range
92: .IB maxupri < pri < maxpri .
93: .TP
94: .BI -R delta
95: Sets the
96: .I run-rate
97: for the share scheduler in seconds.
98: .TP
99: .BI -S maxusers
100: Sets the maximum number of users and groups that can be active.
101: Note that this cannot exceed the maximum configured in the kernel.
102: .TP
103: .BI -U maxusage
104: Upper bound for ``reasonable'' usages.
105: Users with usages larger than this are grouped together and given a
106: .I "normalised usage"
107: which prevents them from interfering with ``normal'' users.
108: .TP
109: .BI -X maxushare
110: If the
111: .SM LIMSHARE
112: scheduling flag is on,
113: then this parameter limits the maximum effective share an
114: individual user can have to
115: .I maxushare
116: times their allocated share.
117: .TP
118: .BI -Y mingshare
119: If the
120: .SM ADJGROUPS
121: scheduling flag is on,
122: and any group is getting less than
123: .I mingshare
124: times its allocated share,
125: the costs incurred by the group members will be adjusted down to compensate.
126: (Does not affect the long-term charges.)
127: .TP
128: .BI -b bio
129: The charge for a disk block I/O operation.
130: .TP
131: .BI -m click
132: The charge for a
133: .IR "memory tick" .
134: .TP
135: .BI -s syscall
136: The charge for a system call.
137: .TP
138: .BI -t tick
139: The charge for a CPU tick.
140: .TP
141: .BI -v
142: Show scheduling feed-back parameters
143: .RI ( charges " only)."
144: .TP
145: .BI -y tio
146: The charge for a stream I/O operation.
147: (This is really dependent on the number of kernel buffer operations,
148: so a
149: .I "write(1)"
150: will cost the same as a
151: .I "write(64)"
152: to an ordinary stream,
153: or a
154: .I "write(1024)"
155: to a pipe.)
156: .TP
157: .IR percent
158: The percentage change to apply to all the charges.
159: .SH EXAMPLES
160: .B charge 10
161: .RS
162: will change the costs to 10% of the default.
163: .RE
164: .B charges
165: .RS
166: show the current charges.
167: .RE
168: .B "charges \-"
169: .RS
170: show the default settings.
171: .RE
172: .SH SEE ALSO
173: .PD 0
174: .TP "\w'/usr/include/sys/charges.hXX'u"
175: .IR /usr/include/sys/charges.h
176: The default values in the kernel.
177: .TP
178: .IR /usr/include/sys/share.h
179: Definition of charges structure.
180: .TP
181: .IR share (5)
182: A description of the
183: .IR "Share Scheduler" .
184: .PD
185: .SH BUGS
186: The percent flag will also affect any new constants,
187: so bias them accordingly.
188: .PP
189: .I Charges
190: works out the current charging percentage
191: by using the difference between the default cost for ``ticks''
192: and the current setting.
193: .PP
194: The defaults are hardly ever relevant.
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.