1QHOST(1) Grid Engine User Commands QHOST(1)
2
6 qhost - show the status of Grid Engine hosts, queues, jobs
7
9 qhost [ -cb ] [ -F [resource_name,...] ] [ -help ] [ -h host_list ] [
10 -j ] [ -l resource=val,... ] [ -u user,... ] [ -xml ].
11
13 qhost shows the current status of the available Grid Engine hosts,
14 queues and the jobs associated with the queues. Selection options allow
15 you to get information about specific hosts, queues, jobs or users. If
16 multiple selections are done a host is only displayed if all selection
17 criteria for a host are met. Without any options qhost will display a
18 list of all hosts without queue or job information.
19
21 -F [ resource_name,... ]
22 qhost will present a detailed listing of the current resource
23 availability per host with respect to all resources (if the
24 option argument is omitted) or with respect to those resources
25 contained in the resource_name list. Please refer to the
26 description of the Full Format in section OUTPUT FORMATS below
27 for further detail.
28 -cb This command line switch can be used since Grid Engine version
30 6.2u5 in combination with other qhost(1) command line switches.
31 In that case the output of the corresponding command will con‐
32 tain information concerning the added job to core binding fea‐
33 ture.
34 If this switch is not used then the command behaves as in ver‐
36 sion 6.2u4 and previous versions.
37 If this option is used then two additional columns will be shown
39 for each displayed host in the output. The first is named NSOC
40 and represents the number of available sockets on that host. The
41 second additional column is named NCOR and it represents the
42 number of cores that are available per socket on the correspond‐
43 ing machine.
44 -help Prints a listing of all options.
46 -h host_list
48 Prints a list of all hosts contained in host_list.
49 -j Prints all jobs running on the queues hosted by the shown hosts.
51 This switch calls -q implicitly.
52 -l resource[=value],...
54 Defines the resources to be granted by the hosts which should be
55 included in the host list output. Matching is performed on hosts
56 based on non-mutable resource availability information only.
57 That means load values are always ignored except the so-called
58 static load values (i.e. "arch", "num_proc", "mem_total",
59 "swap_total" and "virtual_total") ones. Also consumable utiliza‐
60 tion is ignored. If there are multiple -l resource requests
61 they will be concatenated by a logical AND: a host needs to
62 match all resources to be displayed.
63 -q Show information about the queues instances hosted by the dis‐
65 played hosts.
66 -u user,...
68 Display information only on those jobs and queues being associ‐
69 ated with the users from the given user list.
70 -xml This option can be used with all other options and changes the
72 output to XML. The used schemas are referenced in the XML out‐
73 put. The output is printed to stdout.
74 If the -xml parameter is combined with -cb then the XML output
76 will contain additional tags containing information about job to
77 core binding. You can find schema files with the suffix "_cb"
78 in the directory $SGE_ROOT/util/resources/schemas/qhost that
79 describe that changes.
80
82 Depending on the presence or absence of the -q or -F and -j option
83 three output formats need to be differentiated. PP
84 Default Format (without -q, -F and -j)
86 For each host one line is printed. The output consists of consisting of
87 · the Hostname
89 · the Architecture.
91 · the Number of processors.
93 · the Load.
95 · the Total Memory.
97 · the Used Memory.
99 · the Total Swapspace.
101 · the Used Swapspace.
103 If the -q option is supplied, each host status line also contains extra
105 lines for every queue hosted by the host consisting of,
106 · the queue name.
108 · the queue type - one of B(atch), I(nteractive), C(heckpointing),
110 P(arallel), T(ransfer) or combinations thereof,
111 · the number of reserved, used and available job slots,
113 · the state of the queue - one of u(nknown) if the corresponding
115 ge_execd(8) cannot be contacted, a(larm), A(larm), C(alendar sus‐
116 pended), s(uspended), S(ubordinate), d(isabled), D(isabled), E(rror)
117 or combinations thereof.
118 If the state is a(alarm) at least one of the load thresholds defined in
120 the load_thresholds list of the queue configuration (see queue_conf(5))
121 is currently exceeded, which prevents from scheduling further jobs to
122 that queue.
123 As opposed to this, the state A(larm) indicates that at least one of
125 the suspend thresholds of the queue (see queue_conf(5)) is currently
126 exceeded. This will result in jobs running in that queue being succes‐
127 sively suspended until no threshold is violated.
128 The states s(uspended) and d(isabled) can be assigned to queues and
130 released via the qmod(1) command. Suspending a queue will cause all
131 jobs executing in that queue to be suspended.
132 The states D(isabled) and C(alendar suspended) indicate that the queue
134 has been disabled or suspended automatically via the calendar facility
135 of Grid Engine (see calendar_conf(5)), while the S(ubordinate) state
136 indicates, that the queue has been suspend via subordination to another
137 queue (see queue_conf(5) for details). When suspending a queue (regard‐
138 less of the cause) all jobs executing in that queue are suspended too.
139 If an E(rror) state is displayed for a queue, ge_execd(8) on that host
141 was unable to locate the ge_shepherd(8) executable on that host in
142 order to start a job. Please check the error logfile of that
143 ge_execd(8) for leads on how to resolve the problem. Please enable the
144 queue afterwards via the -c option of the qmod(1) command manually.
145 If the -F option was used, resource availability information is printed
147 following the host status line. For each resource (as selected in an
148 option argument to -F or for all resources if the option argument was
149 omitted) a single line is displayed with the following format:
150 · a one letter specifier indicating whether the current resource
152 availability value was dominated by either
153 `g' - a cluster global,
154 `h' - a host total or
155 · a second one letter specifier indicating the source for the current
157 resource availability value, being one of
158 `l' - a load value reported for the resource,
159 `L' - a load value for the resource after administrator defined load
160 scaling has been applied,
161 `c' - availability derived from the consumable resources facility
162 (see complexes(5)),
163 `f' - a fixed availability definition derived from a non-consumable
164 complex attribute or a fixed resource limit.
165 · after a colon the name of the resource on which information is dis‐
167 played.
168 · after an equal sign the current resource availability value.
170 The displayed availability values and the sources from which they
172 derive are always the minimum values of all possible combinations.
173 Hence, for example, a line of the form "qf:h_vmem=4G" indicates that a
174 queue currently has a maximum availability in virtual memory of 4 Giga‐
175 byte, where this value is a fixed value (e.g. a resource limit in the
176 queue configuration) and it is queue dominated, i.e. the host in total
177 may have more virtual memory available than this, but the queue doesn't
178 allow for more. Contrarily a line "hl:h_vmem=4G" would also indicate an
179 upper bound of 4 Gigabyte virtual memory availability, but the limit
180 would be derived from a load value currently reported for the host. So
181 while the queue might allow for jobs with higher virtual memory
182 requirements, the host on which this particular queue resides currently
183 only has 4 Gigabyte available.
184 After the queue status line (in case of -j) a single line is printed
186 for each job running currently in this queue. Each job status line con‐
187 tains
188 · the job ID,
190 · the job name,
192 · the job owner name,
194 · the status of the job - one of t(ransfering), r(unning),
196 R(estarted), s(uspended), S(uspended) or T(hreshold) (see the
197 Reduced Format section for detailed information),
198 · the start date and time and the function of the job (MASTER or SLAVE
200 - only meaningful in case of a parallel job) and
201 · the priority of the jobs.
203
205 GE_ROOT Specifies the location of the Grid Engine standard con‐
206 figuration files.
207 GE_CELL If set, specifies the default Grid Engine cell. To
209 address a Grid Engine cell qhost uses (in the order of
210 precedence):
211 The name of the cell specified in the environment
213 variable GE_CELL, if it is set.
214 The name of the default cell, i.e. default.
216 GE_DEBUG_LEVEL If set, specifies that debug information should be writ‐
219 ten to stderr. In addition the level of detail in which
220 debug information is generated is defined.
221 GE_QMASTER_PORT
223 If set, specifies the tcp port on which ge_qmaster(8) is
224 expected to listen for communication requests. Most
225 installations will use a services map entry for the ser‐
226 vice "sge_qmaster" instead to define that port.
227
229 <ge_root>/<cell>/common/act_qmaster
230 Grid Engine master host file
231
233 ge_intro(1), qalter(1), qconf(1), qhold(1), qmod(1), qstat(1), qsub(1),
234 queue_conf(5), ge_execd(8), ge_qmaster(8), ge_shepherd(8).
235
237 See ge_intro(1) for a full statement of rights and permissions.
238GE 6.2u5 $Date: 2009/11/05 13:18:36 $ QHOST(1)