1PMCD(1)                     General Commands Manual                    PMCD(1)
2
3
4

NAME

6       pmcd - performance metrics collector daemon
7

SYNOPSIS

9       pmcd  [-f]  [-i  ipaddress]  [-l  logfile] [-L bytes] [-n pmnsfile] [-p
10       port[,port ...]  [-q timeout] [-T traceflag] [-t timeout] [-x file]
11

DESCRIPTION

13       pmcd is the collector used by  the  Performance  Co-Pilot  (see  PCPIn‐
14       tro(1))  to  gather  performance metrics on a system.  As a rule, there
15       must be an instance of pmcd running on a  system  for  any  performance
16       metrics to be available to the PCP.
17
18       pmcd accepts connections from client applications running either on the
19       same machine or remotely and  provides  them  with  metrics  and  other
20       related  information  from the machine that pmcd is executing on.  pmcd
21       delegates most of this request servicing to a collection of Performance
22       Metrics Domain Agents (or just agents), where each agent is responsible
23       for a particular group of metrics, known as the domain  of  the  agent.
24       For  example the environ agent is responsible for reporting information
25       relating to the environment of a Challenge system, such as the  cabinet
26       temperature and voltage levels of the power supply.
27
28       The  agents  may be processes started by pmcd, independent processes or
29       Dynamic Shared Objects (DSOs, see dso(5)) attached  to  pmcd's  address
30       space.   The  configuration  section below describes how connections to
31       agents are specified.
32
33       The options to pmcd are as follows.
34
35       -f     By default pmcd is started as a daemon.  The -f option indicates
36              that  it should run in the foreground.  This is most useful when
37              trying to diagnose problems with misbehaving agents.
38
39       -i ipaddress
40              This option is usually only used on hosts  with  more  than  one
41              network  interface.  If no -i options are specified pmcd accepts
42              connections made to any of its  host's  IP  (Internet  Protocol)
43              addresses.   The  -i  option is used to specify explicitly an IP
44              address that  connections  should  be  accepted  on.   ipaddress
45              should  be  in the standard dotted form (e.g. 100.23.45.6).  The
46              -i option may be used multiple times to  define  a  list  of  IP
47              addresses.   Connections made to any other IP addresses the host
48              has will be refused.  This can be used to limit  connections  to
49              one  network  interface if the host is a network gateway.  It is
50              also useful if the host takes over the  IP  address  of  another
51              host  that has failed.  In such a situation only the standard IP
52              addresses of the host should be given (not  the  ones  inherited
53              from  the  failed host).  This allows PCP applications to deter‐
54              mine that a host has failed, rather than connecting to the  host
55              that has assumed the identity of the failed host.
56
57       -l logfile
58              By default a log file named pmcd.log is written in the directory
59              $PCP_LOG_DIR/pmcd.  The -l option causes  the  log  file  to  be
60              written to logfile instead of the default.  If the log file can‐
61              not be created or is not writable,  output  is  written  to  the
62              standard error instead.
63
64       -L bytes
65              PDUs  received by pmcd from monitoring clients are restricted to
66              a maximum size of 65536  bytes  by  default  to  defend  against
67              Denial  of Service attacks.  The -L option may be used to change
68              the maximum incoming PDU size.
69
70       -n pmnsfile
71              Normally pmcd loads the default Performance Metrics  Name  Space
72              (PMNS)  from $PCP_VAR_DIR/pmns/root, however if the -n option is
73              specified an alternative namespace is loaded from the file pmns‐
74              file.
75
76       -q timeout
77              The  pmcd  to  agent version exchange protocol (new in PCP 2.0 -
78              introduced to provide backward compatibility) uses this  timeout
79              to  specify  how  long  pmcd should wait before assuming that no
80              version response is coming from an agent.  If  this  timeout  is
81              reached,  the  agent  is  assumed  to be an agent which does not
82              understand the PCP 2.0 protocol.  The default  timeout  interval
83              is five seconds, but the -q option allows an alternative timeout
84              interval (which must be greater than zero) to be specified.  The
85              unit of time is seconds.
86
87       -t timeout
88              To  prevent  misbehaving  agents from hanging the entire Perfor‐
89              mance Metrics Collection System (PMCS), pmcd  uses  timeouts  on
90              PDU  exchanges with agents running as processes.  By default the
91              timeout interval is five  seconds.   The  -t  option  allows  an
92              alternative  timeout  interval  in  seconds to be specified.  If
93              timeout is zero, timeouts are turned off.  It is almost impossi‐
94              ble  to  use the debugger interactively on an agent unless time‐
95              outs have been turned off for its "parent" pmcd.
96
97              Once pmcd is running, the timeout may be dynamically modified by
98              storing  an integer value (the timeout in seconds) into the met‐
99              ric pmcd.control.timeout via pmstore(1).
100
101       -T traceflag
102              To assist with error diagnosis for agents and/or clients of pmcd
103              that are not behaving correctly, an internal event tracing mech‐
104              anism is supported within  pmcd.   The  value  of  traceflag  is
105              interpreted as a bit field with the following control functions:
106
107              1   enable client connection tracing
108              2   enable PDU tracing
109              256 unbuffered event tracing
110
111              By  default,  event  tracing is buffered using a circular buffer
112              that is over-written as new events are  recorded.   The  default
113              buffer  size  holds the last 20 events, although this number may
114              be  over-ridden  by  using  pmstore(1)  to  modify  the   metric
115              pmcd.control.tracebufs.
116
117              Similarly once pmcd is running, the event tracing control may be
118              dynamically modified by storing 1 (enable) or 0  (disable)  into
119              the  metrics  pmcd.control.traceconn,  pmcd.control.tracepdu and
120              pmcd.control.tracenobuf.  These metrics map to  the  bit  fields
121              associated with the traceflag argument for the -T option.
122
123              When  operating in buffered mode, the event trace buffer will be
124              dumped whenever an agent connection is terminated  by  pmcd,  or
125              when  any value is stored into the metric pmcd.control.dumptrace
126              via pmstore(1).
127
128              In unbuffered mode, every event will be reported when it occurs.
129
130       -x file
131              Before the pmcd logfile can be  opened,  pmcd  may  encounter  a
132              fatal  error  which  prevents it from starting.  By default, the
133              output describing this error is sent  to  /dev/tty  but  it  may
134              redirected to file.
135
136       If  a  PDU exchange with an agent times out, the agent has violated the
137       requirement that it delivers metrics with little or no delay.  This  is
138       deemed a protocol failure and the agent is disconnected from pmcd.  Any
139       subsequent requests for information from the agent  will  fail  with  a
140       status indicating that there is no agent to provide it.
141
142       It  is  possible  to  specify  host-level access control to pmcd.  This
143       allows one to prevent users from certain hosts from accessing the  met‐
144       rics provided by pmcd and is described in more detail in the Section on
145       ACCESS CONTROL below.
146

CONFIGURATION

148       On startup  pmcd  looks  for  a  configuration  file  named  $PCP_PMCD‐
149       CONF_PATH.   This  file  specifies which agents cover which performance
150       metrics domains and how pmcd should make contact with the  agents.   An
151       optional  section  specifying host-based access controls may follow the
152       agent configuration data.
153
154       Warning: pmcd is usually started as part of the boot sequence and  runs
155       as  root.   The configuration file may contain shell commands to create
156       agents, which will be executed by root.  To prevent  security  breaches
157       the  configuration  file  should  be writable only by root.  The use of
158       absolute path names is also recommended.
159
160       The case of the reserved words in the configuration  file  is  unimpor‐
161       tant, but elsewhere, the case is preserved.
162
163       Blank lines and comments are permitted (even encouraged) in the config‐
164       uration file.  A comment begins with a ``#'' character and finishes  at
165       the end of the line.  A line may be continued by ensuring that the last
166       character on the line is a ``\'' (backslash).  A comment on a continued
167       line  ends at the end of the continued line.  Spaces may be included in
168       lexical elements by enclosing  the  entire  element  in  double  quotes
169       (there  must  be  whitespace  before  the opening and after the closing
170       quote).  A double quote preceded by a backslash  is  always  a  literal
171       double  quote.   A ``#'' in double quotes or preceded by a backslash is
172       treated literally rather than as a comment delimiter.  Lexical elements
173       and separators are described further in the following sections.
174

AGENT CONFIGURATION

176       Each  line of the agent configuration section of the configuration file
177       contains details of how to connect pmcd to one of its agents and speci‐
178       fies  which  metrics  domain  the  agent  deals  with.  An agent may be
179       attached as a DSO, or via a socket, or a pair of pipes.
180
181       Each line of the agent configuration section of the configuration  file
182       must  be  either  an  agent  specification, a comment, or a blank line.
183       Lexical elements are separated by whitespace characters, however a sin‐
184       gle  agent  specification  may  not  be  broken across lines unless a \
185       (backslash) is used to continue the line.
186
187       Each agent specification must start with a textual label (string)  fol‐
188       lowed  by an integer in the range 1 to 510.  The label is a tag used to
189       refer to the agent and the integer specifies the domain for  which  the
190       agent  supplies data.  This domain identifier corresponds to the domain
191       portion of the PMIDs handled by the agent.   Each  agent  must  have  a
192       unique label and domain identifier.
193
194       For DSO agents a line of the form:
195
196              label domain-no dso entry-point path
197
198       should appear.  Where,
199
200       label         is a string identifying the agent
201       domain-no     is  an  unsigned integer specifying the agent's domain in
202                     the range 1 to 510
203       entry-point   is the name of an initialization function which  will  be
204                     called when the DSO is loaded
205       path          designates the location of the DSO. This field is treated
206                     differently on Irix and on Linux. Later expects it to  be
207                     an  absolute  pathname, while former uses some heuristics
208                     to find an agent. If path begins with a / it is taken  as
209                     an absolute path specifying the DSO. If path is relative,
210                     pmcd will expect to find the agent in  a  file  with  the
211                     name mips_simabi.path, where simabi is either o32, n32 or
212                     64.  pmcd is only able to load DSO agents that  have  the
213                     same  simabi  (Subprogram Interface Model ABI, or calling
214                     conventions) as it does (i.e. only one of the simabi ver‐
215                     sions  will be applicable).  The simabi version of a run‐
216                     ning pmcd may  be  determined  by  fetching  pmcd.simabi.
217                     Alternatively,  the file(1) command may be used to deter‐
218                     mine the simabi version from the pmcd executable.
219
220                     For a relative path the  environment  variable  PMCD_PATH
221                     defines  a  colon  (:)  separated  list of directories to
222                     search when trying to locate the agent DSO.  The  default
223                     search path is $PCP_SHARE_DIR/lib:/usr/pcp/lib.
224
225       For agents providing socket connections, a line of the form
226
227              label domain-no socket addr-family address [ command ]
228
229       should appear.  Where,
230
231       label         is a string identifying the agent
232       domain-no     is  an  unsigned integer specifying the agent's domain in
233                     the range 1 to 510
234       addr-family   designates whether  the  socket  is  in  the  AF_INET  or
235                     AF_UNIX  domain,  and  the  corresponding values for this
236                     parameter are inet and unix respectively.
237       address       specifies the address of the socket within the previously
238                     specified  addr-family.   For  unix  sockets, the address
239                     should be the name of an agent's socket on the local host
240                     (a valid address for the UNIX domain).  For inet sockets,
241                     the address may be either a port number or  a  port  name
242                     which  may  be  used  to connect to an agent on the local
243                     host.  There is no syntax for specifying an  agent  on  a
244                     remote  host as a pmcd deals only with agents on the same
245                     machine.
246       command       is an optional parameter used to specify a  command  line
247                     to  start the agent when pmcd initializes.  If command is
248                     not present, pmcd assumes that the  specified  agent  has
249                     already been created.  The command is considered to start
250                     from the  first  non-white  character  after  the  socket
251                     address  and  finish  at the next newline that isn't pre‐
252                     ceded by a backslash.  After a  fork(2)  the  command  is
253                     passed unmodified to execve(2) to instantiate the agent.
254
255       For  agents  interacting  with the pmcd via stdin/stdout, a line of the
256       form:
257
258              label domain-no pipe protocol command
259
260       should appear.  Where,
261
262       label         is a string identifying the agent
263       domain-no     is an unsigned integer specifying the agent's domain
264       protocol      The value for this parameter should be binary.
265
266                     Additionally, the protocol can include the notready  key‐
267                     word  to  indicate  that  the agent must be marked as not
268                     being ready to process requests from pmcd. The agent will
269                     explictily  notify  the  pmcd when it is ready to process
270                     the requests by sending PM_ERR_PMDAREADY PDU.
271
272       command       specifies a command line to start  the  agent  when  pmcd
273                     initializes.   Note  that  command is mandatory for pipe-
274                     based agents.  The command is considered  to  start  from
275                     the  first non-white character after the protocol parame‐
276                     ter and finish at the next newline that isn't preceded by
277                     a  backslash.   After  a  fork(2)  the  command is passed
278                     unmodified to execve(2) to instantiate the agent.
279

ACCESS CONTROL CONFIGURATION

281       The access control section of the configuration file is  optional,  but
282       if  present  it  must follow the agent configuration data.  The case of
283       reserved words is ignored, but elsewhere case  is  preserved.   Lexical
284       elements  in  the access control section are separated by whitespace or
285       the special delimiter characters: square brackets  (``[''  and  ``]''),
286       braces  (``{''  and  ``}''), colon (``:''), semicolon (``;'') and comma
287       (``,'').  The special characters are not  treated  as  special  in  the
288       agent configuration section.
289
290       The  access  control  section of the file must start with a line of the
291       form:
292
293       [access]
294
295       Leading and trailing whitespace may appear around and within the brack‐
296       ets  and  the case of the access keyword is ignored.  No other text may
297       appear on the line except a trailing comment.
298
299       Following this line, the remainder of  the  configuration  file  should
300       contain  lines  that allow or disallow operations from particular hosts
301       or groups of hosts.
302
303       There are two kinds of operations that occur via pmcd:
304
305       fetch          allows retrieval of information from pmcd.  This may  be
306                      information   about  a  metric  (e.g.  its  description,
307                      instance domain or help text) or a value for a metric.
308
309       store          allows pmcd to be used to store metric values in  agents
310                      that permit store operations.
311
312       Access  to  pmcd is granted at the host level, i.e. all users on a host
313       are granted the same level of access.  Permission to perform the  store
314       operation should not be given indiscriminately; it has the potential to
315       be abused by malicious users.
316
317       Hosts may be identified by name, IP address or a wildcarded IP  address
318       with the single wildcard character ``*'' as the last-given component of
319       the IP address.  Host names may not be wildcarded.  The  following  are
320       all valid host identifiers:
321
322            boing
323            localhost
324            giggle.melbourne.sgi.com
325            129.127.112.2
326            129.127.114.*
327            129.*
328            *
329
330       The following are not valid host identifiers:
331
332            *.melbourne
333            129.127.*.*
334            129.*.114.9
335            129.127*
336
337       The  first  example  is not allowed because only (numeric) IP addresses
338       may contain a wildcard.  The second example is not valid because  there
339       is  more  than  one wildcard character.  The third contains an embedded
340       wildcard, the fourth has a wildcard character that is not the last com‐
341       ponent of the IP address (the last component is 127*).
342
343       The  name  localhost is given special treatment to make the behavior of
344       host wildcarding consistent.  Rather than being 127.0.0.1, it is mapped
345       to the primary IP address associated with the name of the host on which
346       pmcd is running.  Beware of  this  when  running  pmcd  on  multi-homed
347       hosts.
348
349       Access  for hosts are allowed or disallowed by specifying statements of
350       the form:
351
352              allow hostlist : operations ;
353              disallow hostlist : operations ;
354
355       hostlist      is a comma separated list of host identifiers.
356
357       operations    is  a  comma  separated  list  of  the  operation   types
358                     described  above,  all (which allows/disallows all opera‐
359                     tions), or all except operations (which  allows/disallows
360                     all operations except those listed).
361
362       Where  no  specific allow or disallow statement applies to an operation
363       for some host, the default is to allow the operation  from  that  host.
364       In the trivial case when there is no access control section in the con‐
365       figuration file, all operations from all hosts are permitted.
366
367       If a new connection to pmcd is attempted from a host that is  not  per‐
368       mitted to perform any operations, the connection will be closed immedi‐
369       ately after an error response PM_ERR_PERMISSION has been  sent  to  the
370       client attempting the connection.
371
372       Statements  with  the  same  level  of wildcarding specifying identical
373       hosts may not contradict each other.  For example if a host named clank
374       had  an IP address of 129.127.112.2, specifying the following two rules
375       would be erroneous:
376
377            allow clank : fetch, store;
378            disallow 129.127.112.2 : all except fetch;
379
380       because they both refer to the same host, but disagree  as  to  whether
381       the fetch operation is permitted from that host.
382
383       Statements  containing  more specific host specifications override less
384       specific ones according to the level of  wildcarding.   For  example  a
385       rule of the form
386
387            allow clank : all;
388
389       overrides
390
391            disallow 129.127.112.* : all except fetch;
392
393       because the former contains a specific host name (equivalent to a fully
394       specified IP address), whereas the latter has a wildcard.  In turn, the
395       latter would override
396
397            disallow * : all;
398
399       It  is possible to limit the number of connections from a host to pmcd.
400       This may be done by adding a clause of the form
401
402              maximum n connections
403
404       to the operations list of an allow statement.  Such a clause may not be
405       used in a disallow statement.  Here, n is the maximum number of connec‐
406       tions that will be accepted from hosts matching the host  identifier(s)
407       used in the statement.
408
409       An  access control statement with a list of host identifiers is equiva‐
410       lent to a group of access control statements, with each specifying  one
411       of  the  host identifiers in the list and all with the same access con‐
412       trols (both permissions and connection limits).  A wildcard  should  be
413       used if you want hosts to contribute to a shared connection limit.
414
415       When  a  new client requests a connection, and pmcd has determined that
416       the client has permission to connect, it searches the matching list  of
417       access control statements for the most specific match containing a con‐
418       nection limit.  For brevity, this will be called  the  limiting  state‐
419       ment.   If there is no limiting statement, the client is granted a con‐
420       nection.  If there is a limiting  statement  and  the  number  of  pmcd
421       clients  with IP addresses that match the host identifier in the limit‐
422       ing statement is less than the connection limit in the  statement,  the
423       connection is allowed.  Otherwise the connection limit has been reached
424       and the client is refused a connection.
425
426       The wildcarding in host  identifiers  means  that  once  pmcd  actually
427       accepts  a  connection  from a client, the connection may contribute to
428       the current connection count of more than one access control  statement
429       (the  client's  host may match more than one access control statement).
430       This may be significant for subsequent connection requests.
431
432       Note that because most specific match semantics are used when  checking
433       the  connection  limit, priority is given to clients with more specific
434       host identifiers.  It is also possible to exceed connection  limits  in
435       some situations.  Consider the following:
436
437              allow clank : all, maximum 5 connections;
438              allow * : all except store, maximum 2 connections;
439
440       This  says  that  only 2 client connections at a time are permitted for
441       all hosts other than "clank", which is permitted 5.  If a  client  from
442       host "boing" is the first to connect to pmcd, its connection is checked
443       against the second statement (that is the most specific  match  with  a
444       connection  limit).   As  there are no other clients, the connection is
445       accepted and contributes towards the limit for only the  second  state‐
446       ment  above.   If the next client connects from "clank", its connection
447       is checked against the limit for the first  statement.   There  are  no
448       other  connections  from  "clank", so the connection is accepted.  Once
449       this connection is accepted, it counts towards both statements'  limits
450       because "clank" matches the host identifier in both statements.  Remem‐
451       ber that the decision to accept a new connection is made using only the
452       most  specific  matching  access  control  statement  with a connection
453       limit.  Now, the connection limit for the  second  statement  has  been
454       reached.   Any  connections  from  hosts  other  than  "clank"  will be
455       refused.
456
457       If instead, pmcd with  no  clients  saw  three  successive  connections
458       arrived  from  "boing",  the  first two would be accepted and the third
459       refused.  After that, if a connection was  requested  from  "clank"  it
460       would  be accepted.  It matches the first statement, which is more spe‐
461       cific than the second, so the connection limit in the first is used  to
462       determine  that  the  client has the right to connect.  Now there are 3
463       connections contributing to the second  statement's  connection  limit.
464       Even  though  the  connection  limit  for the second statement has been
465       exceeded, the earlier connections from  "boing"  are  maintained.   The
466       connection  limit  is only checked at the time a client attempts a con‐
467       nection rather than being re-evaluated every time a new client connects
468       to pmcd.
469
470       This gentle scheme is designed to allow reasonable limits to be imposed
471       on a first come first served basis, with specific exceptions.
472
473       As illustrated by the example above, a client's connection  is  honored
474       once it has been accepted.  However, pmcd reconfiguration (see the next
475       section) re-evaluates all the connection counts and will  cause  client
476       connections to be dropped where connection limits have been exceeded.
477

RECONFIGURING PMCD

479       If  the  configuration  file  has  been  changed  or if an agent is not
480       responding because it has terminated or the PMNS has been changed, pmcd
481       may be reconfigured by sending it a SIGHUP, as in
482
483            # pmsignal -a -s HUP pmcd
484
485       When  pmcd  receives  a  SIGHUP,  it  checks the configuration file for
486       changes.  If the file has been modified, it is reparsed  and  the  con‐
487       tents become the new configuration.  If there are errors in the config‐
488       uration file, the existing configuration is retained and  the  contents
489       of the file are ignored.  Errors are reported in the pmcd log file.
490
491       It  also  checks  the  PMNS file for changes. If the PMNS file has been
492       modified, then it is reloaded.  Use of tail(1) on the log file is  rec‐
493       ommended while reconfiguring pmcd.
494
495       If the configuration for an agent has changed (any parameter except the
496       agent's label is different), the agent is restarted.  Agents whose con‐
497       figurations  do  not change are not restarted.  Any existing agents not
498       present in the new configuration are terminated.  Any  deceased  agents
499       are that are still listed are restarted.
500
501       Sometimes  it  is  necessary to restart an agent that is still running,
502       but malfunctioning.  Simply stop the agent  (e.g.  using  SIGTERM  from
503       pmsignal(1)), then send pmcd a SIGHUP, which will cause the agent to be
504       restarted.
505

STARTING AND STOPPING PMCD

507       Normally, pmcd is started automatically at boot time and  stopped  when
508       the system is being brought down (see rc2(1M) and rc0(1M)).  Under cer‐
509       tain circumstances it is necessary to start or stop pmcd manually.   To
510       do this one must become superuser and type
511
512            # $PCP_RC_DIR/pcp start
513
514       to start pmcd, or
515
516            # $PCP_RC_DIR/pcp stop
517
518       to  stop pmcd.  Starting pmcd when it is already running is the same as
519       stopping it and then starting it again.
520
521       Sometimes it may be necessary to restart pmcd during another  phase  of
522       the  boot  process.  Time-consuming parts of the boot process are often
523       put into the background to allow the system to become available  sooner
524       (e.g.  mounting huge databases).  If an agent run by pmcd requires such
525       a task to complete before it can  run  properly,  it  is  necessary  to
526       restart  or  reconfigure  pmcd after the task completes.  Consider, for
527       example, the case of mounting a database in the background while  boot‐
528       ing.   If the PMDA which provides the metrics about the database cannot
529       function until the database  is  mounted  and  available  but  pmcd  is
530       started  before the database is ready, the PMDA will fail (however pmcd
531       will still service requests for metrics from other  domains).   If  the
532       database is initialized by running a shell script, adding a line to the
533       end of the script to reconfigure pmcd (by sending  it  a  SIGHUP)  will
534       restart the PMDA (if it exited because it couldn't connect to the data‐
535       base).  If the PMDA didn't exit in such a situation it would be  neces‐
536       sary  to  restart pmcd because if the PMDA was still running pmcd would
537       not restart it.
538
539       Normally pmcd listens for client connections on one or more  well-known
540       TCP/IP port numbers (historically 4321 and more recently the officially
541       registered port 44321; in the current release,  pmcd  listens  on  only
542       port  44321  by default).  Either the environment variable PMCD_PORT or
543       the -p command line option may be used to specify alternative port num‐
544       ber(s)  when  pmcd  is  started;  in  each case, the specficiation is a
545       comma-separated list of one or more  numerical  port  numbers.   Should
546       both methods be used or multiple -p options appear on the command line,
547       pmcd will listen on the union of the set of ports specified via all  -p
548       options  and  the PMCD_PORT environment variable.  If non-default ports
549       are used with pmcd care should be taken to  ensure  that  PMCD_PORT  is
550       also set in the environment of any client application that will connect
551       to pmcd.
552

FILES

554       $PCP_PMCDCONF_PATH
555                 default configuration file
556       $PCP_PMCDOPTIONS_PATH
557                 command   line   options   to   pmcd   when   launched   from
558                 $PCP_RC_DIR/pcp  All  the  command  line  option lines should
559                 start with a hyphen as the first character.   This  file  can
560                 also contain environment variable settings of the form "VARI‐
561                 ABLE=value".
562       ./pmcd.log
563                 (or $PCP_LOG_DIR/pmcd/pmcd.log when started automatically)
564       $PCP_RUN_DIR/pmcd.pid
565                 contains an ascii decimal representation of the process ID of
566                 pmcd , when it's running.
567                 All messages and diagnostics are directed here
568

ENVIRONMENT

570       In addition to the PCP environment variables described in the PCP ENVI‐
571       RONMENT section below, the PMCD_PORT variable is also recognised as the
572       TCP/IP port for incoming connections (default 44321).
573

PCP ENVIRONMENT

575       Environment variables with the prefix PCP_ are used to parameterize the
576       file and directory names used by PCP.  On each installation,  the  file
577       /etc/pcp.conf  contains  the  local  values  for  these variables.  The
578       $PCP_CONF variable may be used to specify an alternative  configuration
579       file, as described in pcp.conf(4).
580

SEE ALSO

582       PCPIntro(1),  pmdbg(1),  pmerr(1),  pmgenmap(1),  pminfo(1), pmstat(1),
583       pmstore(1), pmval(1), pcp.conf(4), and pcp.env(4).
584

DIAGNOSTICS

586       If pmcd is already running the message "Error: OpenRequestSocket  bind:
587       Address  may  already  be in use" will appear.  This may also appear if
588       pmcd was shutdown with an outstanding request from a client.   In  this
589       case,  a  request socket has been left in the TIME_WAIT state and until
590       the system closes it down (after some timeout period) it  will  not  be
591       possible to run pmcd.
592
593       In  addition  to  the  standard PCP debugging flags, see pmdbg(1), pmcd
594       currently uses DBG_TRACE_APPL0  for  tracing  I/O  and  termination  of
595       agents, DBG_TRACE_APPL1 for tracing host access control (see below) and
596       DBG_TRACE_APPL2 for tracing the configuration file scanner and parser.
597

CAVEATS

599       pmcd does not explicitly  terminate  its  children  (agents),  it  only
600       closes  their pipes.  If an agent never checks for a closed pipe it may
601       not terminate.
602
603       The configuration file parser will only read lines of  less  than  1200
604       characters.  This is intended to prevent accidents with binary files.
605
606       The  timeouts controlled by the -t option apply to IPC between pmcd and
607       the PMDAs it spawns.  This is independent of settings of  the  environ‐
608       ment   variables  PMCD_CONNECT_TIMEOUT  and  PMCD_REQUEST_TIMEOUT  (see
609       PCPIntro(1)) which may be used respectively  to  control  timeouts  for
610       client  applications  trying  to  connect to pmcd and trying to receive
611       information from pmcd.
612
613
614
615Performance Co-Pilot                  SGI                              PMCD(1)
Impressum