1PMCONTROLLOG(3) Library Functions Manual PMCONTROLLOG(3)
2
6 __pmControlLog - enable, disable or enquire about logging of perfor‐
7 mance metrics
8
10 #include <pcp/pmapi.h>
11 #include <pcp/impl.h>
12 int __pmControlLog(int fd, const pmResult *request, int control, int state, int delta, pmResult **status)
14 cc ... -lpcp
16
18 __pmControlLog may be used to enable or disable the archive logging for
19 particular performance metrics, as identified by the request parameter;
20 see pmFetch(3) for an explanation of the pmResult structure.
21 The application must have previously issued a call to __pmConnectLog‐
23 ger(3) to establish a control-port connection to the pmlogger(1)
24 instance to whom the control request is to be directed, and fd (the
25 result from __pmConnectLogger(3)) identifies this connection.
26 Within request, only the details of the performance metrics and their
28 associated instances will be used, i.e. the values of the metrics, if
29 any, will be ignored. request would typically be constructed as the
30 result of an earlier call to pmFetch(3). For metrics with a singular
31 value (having an instance domain of PM_INDOM_NULL) the corresponding
32 pmValueSet should have the value one in the numval field and PM_IN_NULL
33 as the inst field of the single pmValue supplied. If multiple explicit
34 instances are to be logged, the numval field of the pmValueSet should
35 contain the number of instances supplied and the inst fields of the
36 pmValue structures should contain specific instance identifiers (which
37 may not have the reserved value PM_IN_NULL).
38 If the numval field within any of the pmValueSet structures in request
40 has a value of zero, it indicates that all available instances of the
41 metric should be used. Enumeration of the instance domain is deferred
42 until the logger fetches the metric prior to writing it to the log,
43 rather than being performed when the __pmControlLog request is
44 received. This is useful for metrics with instance domains that change
45 over time. It is an error to specify numval equal to zero if the cor‐
46 responding metric has a singular value (no instance domain).
47 There are several sorts of logging control available, namely mandatory
49 or advisory, as defined by the control argument, and on, off or maybe
50 as defined by the state argument. These different types of control may
51 be used to ensure that some performance metrics can be guaranteed to
52 always be in the log, while others may be dynamically enabled or dis‐
53 abled as determined by the level and type of system activity.
54 The actual action to be performed is defined by the combination of con‐
56 trol and state as follows. If control is PM_LOG_MANDATORY and state is
57 PM_LOG_ON, then logging is enabled. If control is PM_LOG_MANDATORY and
58 state is PM_LOG_OFF, then logging is disabled. If control is
59 PM_LOG_MANDATORY and state is PM_LOG_MAYBE, then subsequent advisory
60 controls will be honored. If the logging state prior to the request
61 was mandatory (on or off), the state is changed to advisory off. If
62 the logging state was already advisory (either on or off), it remains
63 unchanged. If control is PM_LOG_ADVISORY and the last mandatory con‐
64 trol for the metric was PM_LOG_MAYBE, then logging is enabled or dis‐
65 abled as specified by the state argument, i.e. PM_LOG_ON or
66 PM_LOG_OFF. When the arguments state and control specify a request to
67 change the logging behavior, the argument delta defines the logging
68 interval in milliseconds to be applied to all metrics and instances
69 identified in request.
70 The result argument status returns the current logging state for each
72 of the nominated performance metrics. There is a 1:1 correspondence
73 between the elements of request and status. For metrics in request
74 that have pmValueSets with numval equal to zero, the corresponding
75 pmValueSet in result will contain a value for each available instance
76 at the time of the call. Each metric value in status will have the
77 current logging state encoded in it. The detailed outcome of the oper‐
78 ation for each metric can be determined by comparing these values to
79 that requested via control, state and delta.
80 Macros defined in <pcp/impl.h> may be used to extract the state and
82 logging interval from the returned metric values. PMLC_GET_ON returns
83 true if logging is on, or false if it is off; PMLC_GET_MAND returns
84 true if logging is mandatory, or false if it is advisory;
85 PMLC_GET_INLOG returns true if the metric has been logged at least
86 once, or false otherwise; PMLC_GET_AVAIL returns true if the metric was
87 available from its source the last time it was supposed to be logged,
88 or false if it was unavailable; and PMLC_GET_DELTA returns the current
89 logging interval for the metric (in milliseconds). PMLC_MAX_DELTA
90 defines the greatest delta that can be returned in an encoded metric
91 value.
92 As a special case, when control is PM_LOG_ENQUIRE, state and delta are
94 ignored, and status returns the current logging state of the nominated
95 performance metrics (this variant makes no changes to the logging
96 state).
97 If the value of the logging interval is 0, either for delta in a
99 request to change state to PM_LOG_ON, or encoded in the value returned
100 from PM_LOG_ENQUIRE, then this corresponds to the special ``once only''
101 logging of metrics that appear once in the archive log, and are never
102 logged again.
103 __pmControlLog returns zero on success.
105
107 pmlc(1), pmlogger(1), PMAPI(3), pmFetch(3) and __pmConnectLogger(3).
108
110 PM_ERR_TOOSMALL
111 The number of metrics in request is less than one.
112 PM_ERR_VALUE
114 One or more of the pmValueSets in request had numval (the number
115 of instances) less than one.
116 EINVAL An invalid combination of control and state was specified, or
118 delta was negative.
119Performance Co-Pilot SGI PMCONTROLLOG(3)