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 Univa Grid Engine
service daemon. The current Univa Grid Engine implementation allows one
to query the SGE_QMASTER daemon and any running SGE_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 Univa 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 mes-
sages 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.
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 applica-
tion 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 Univa Grid Engine dae-
mons:
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
execd:
The info message contains information for the execd job dis-
patcher:
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 <sge_root>/<cell>/com-
mon/host_aliases. If this option is used it is not necessary to set
any Univa Grid Engine environment variable.
-ssl
This option can be used to specify an SSL (Secure Socket Layer) config-
uration. 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
-dump
This option allows an administrator to observe the communication proto-
col data flow of a Univa 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 envi-
ronment variable can be set to any combination of the following speci-
fiers 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 con-
tacted 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 sge_qmaster/sge_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).
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
SGE_ROOT Specifies the location of the Univa Grid Engine standard
configuration files.
SGE_CELL If set, specifies the default Univa Grid Engine cell.
SEE ALSO
sge_intro(1), host_aliases(5), sge_qmaster(8), sge_execd(8).
COPYRIGHT
See sge_intro(1) for a full statement of rights and permissions.
UGE 8.0.0 $Date: 2009/03/12 16:06:25 $ QPING(1)
Man(1) output converted with
man2html