qping

Langue: en

Autres versions - même langue

Version: $Date: 2008/04/30 09:11:03 $ (fedora - 04/07/09)

Section: 1 (Commandes utilisateur)

NAME

qping - check application status of Grid Engine daemons.

SYNTAX

qping [-help] [-noalias] [-ssl|-tcp] [ [ [-i <interval>] [-info] [-f] ] | [ [-dump_tag tag [param] ] [-dump] [-nonewline] ] ] <host> <port> <name> <id>

DESCRIPTION

Qping is used to validate the runtime status of a Grid Engine service daemon. The current Grid Engine implementation allows one to query the GE_QMASTER daemon and any running GE_EXECD daemon. The qping command is used to send a SIM (Status Information Message) to the destination daemon. The communication layer of the specified daemon will respond with a SIRM (Status Information Response Message) which contains status information about the consulted daemon.

The qping -dump and -dump_tag options allowing an administrator to observe the communication protocol data flow of a Grid Engine service daemon. The qping -dump instruction must be started with root account and on the same host where the observed daemon is running.

OPTIONS

-f

Show full status information on each ping interval.

First output Line: The first output line shows the date and time of the request.

SIRM version: Internal version number of the SIRM (Status Information Response Message)

SIRM message id: Current message id for this connection

start time: Start time of daemon. The format is as follows:

MM/DD/YYYY HH:MM:SS (seconds since 01.01.1970)

run time [s]: Run time in seconds since start time

messages in read buffer: Nr. of buffered messages in communication buffer. The messages are buffered for the application (daemon). When this number grows too large the daemon is not able to handle all messages sent to it.

messages in write buffer: Nr. of buffered messages in the communication write buffer. The messages are sent from the application (daemon) to the connected clients, but the communication layer wasn't able to send the messages yet. If this number grows too large, the communication layer isn't able to send them as fast as the application (daemon) wants the messages to be sent.

nr. of connected clients: This is the number of actual connected clients to this daemon. This also implies the current qping connection.

status: The status value of the daemon. This value depends on the application which reply to the qping request. If the application does not provide any information the status is 99999. Here are the possible status information values for the Grid Engine daemons:

qmaster:

0 There is no unusual timing situation.

1 One or more threads has reached warning timeout. This may happen when at least one thread does not increment his time stamp for a not usual long time. A possible reason for this is a high workload for this thread.
     2 One or more threads has reached error timeout. This may happen when at least one thread has not incremented his time stamp for longer than 10 minutes.

3 The time measurement is not initialized.

execd:

0 There is no unusual timing situation.

1 Dispatcher has reached warning timeout. This may happen when the dispatcher does not increment his time stamp for a unusual long time. A possible reason for this is a high workload.
     2 Dispatcher has reached error timeout. This may happen when the dispatcher has not incremented his time stamp for longer than 10 minutes.

3 The time measurement is not initialized.

info: Status message of the daemon. This value depends on the application which reply to the qping request. If the application does not provide any information the info message is "not available". Here are the possible status information values for the Grid Engine daemons:

qmaster:

The info message contains information about the qmaster threads followed by a thread state and time information. Each time when one of the known threads pass through their main loop the time information is updated. Since the qmaster has two message threads every message thread updates the time. This means the timeout for the message thread (MT) can only occur when no message thread is active anymore:

THREAD_NAME: THREAD_STATE (THREAD_TIME)

THREAD_NAME:

 MAIN: Main thread 
 signaler: Signal thread 
 event_master: Event master thread 
 timer: Timer thread 
 worker: Worker thread
 listener: Listener thread
 scheduler: Scheduler thread
 jvm: Java thread 
 
 The thread names above will be followed by a 3 digit number.
 

THREAD_STATE:

 R: Running
 W: Warning
 E: Error
 

THREAD_TIME:

 Time since last timestamp updating.
 

After the dispatcher information follows an additional information string which describes the complete application status.

execd:

The info message contains information for the execd job dispatcher: dispatcher: STATE (TIME)

STATE:

 R: Running
 W: Warning
 E: Error
 

TIME:

 Time since last timestamp updating.
 

After the thread information follows an additional information string which describes the application status.

Monitor: If available, displays statistics on a thread. The data for each thread is displayed in one line. The format of this line can be changed at any time. Only the master implements the monitoring.

-help

Prints a list of all options.

-i interval

Set qping interval time.

The default interval time is one second. Qping will send a SIM (Status Information Message) on each interval time.

-info

Show full status information (see -f for more information) and exit. The exit value 0 indicates no error. On errors qping returns with 1.

-noalias

Ignore host_aliases file, which is located at <ge_root>/<cell>/common/host_aliases. If this option is used it is not necessary to set any Grid Engine environment variable.

-ssl

This option can be used to specify an SSL (Secure Socket Layer) configuration. The qping will use the configuration to connect to services running SSL. If the SGE settings file is not sourced, you have to use the -noalias option to bypass the need for the SGE_ROOT environment variable. The following environment variables are used to specify your certificates:
  SSL_CA_CERT_FILE - CA certificate file
  SSL_CERT_FILE    - certificates file
  SSL_KEY_FILE     - key file
  SSL_RAND_FILE    - rand file

-tcp

This option is used to select TCP/IP as the protocol used to connect to other services.

-nonewline

Dump output will not have a linebreak within a message and binary messages are not unpacked.

-dump

This option allows an administrator to observe the communication protocol data flow of a Grid Engine service daemon. The qping -dump instruction must be started as root and on the same host where the observed daemon is running.

The output is written to stdout. The environment variable "SGE_QPING_OUTPUT_FORMAT" can be set to hide columns, set a default column width or to set a hostname output format. The value of the environment variable can be set to any combination of the following specifiers separated by a space character:

 "h:X"   -> hide column X
 "s:X"   -> show column X
 "w:X:Y" -> set width of column X to Y
 "hn:X"  -> set hostname output parameter X. 
            X values are "long" or "short"
 

Start qping -help to see which columns are available.

-dump_tag tag [param]

This option has the same the same meaning as -dump, but can provide more information by specifying the debug level and message types qping should print: -dump_tag ALL <debug level> This option shows all possible debug messages (APP+MSG) for the debug levels, ERROR, WARNING, INFO, DEBUG and DPRINTF. The contacted service must support this kind of debugging. This option is not currently implemented. -dump_tag APP <debug level> This option shows only application debug messages for the debug levels, ERROR, WARNING, INFO, DEBUG and DPRINTF. The contacted service must support this kind of debugging. This option is not currently implemented. -dump_tag MSG This option has the same behavior as the -dump option.

host

Host where daemon is running.

port

Port which daemon has bound (used ge_qmaster/ge_execd port number).

name

Name of communication endpoint ("qmaster" or "execd"). A communication endpoint is a triplet of hostname/endpoint name/endpoint id (e.g. hostA/qmaster/1 or subhost/qstat/4).

id

Id of communication endpoint ("1" for daemons)

EXAMPLES

 >qping master_host 31116 qmaster
 08/24/2004 16:41:15 endpoint master_host/qmaster/1 at port 31116 is up since 365761 seconds
 08/24/2004 16:41:16 endpoint master_host/qmaster/1 at port 31116 is up since 365762 seconds
 08/24/2004 16:41:17 endpoint master_host/qmaster/1 at port 31116 is up since 365763 seconds
 
 > qping -info master_host 31116 qmaster 1
 08/24/2004 16:42:47:
 SIRM version:             0.1
 SIRM message id:          1
 start time:               08/20/2004 11:05:14 (1092992714)
 run time [s]:             365853
 messages in read buffer:  0
 messages in write buffer: 0
 nr. of connected clients: 4
 status:                   0
 info:                     ok
 
 > qping -info execd_host 31117 execd 1
 08/24/2004 16:43:45:
 SIRM version:             0.1
 SIRM message id:          1
 start time:               08/20/2004 11:06:13 (1092992773)
 run time [s]:             365852
 messages in read buffer:  0
 messages in write buffer: 0
 nr. of connected clients: 2
 status:                   0
 info:                     ok
 

ENVIRONMENTAL VARIABLES

GE_ROOT
Specifies the location of the Grid Engine standard configuration files.
GE_CELL
If set, specifies the default Grid Engine cell.

SEE ALSO

See for a full statement of rights and permissions.