1SGE_CONF(5) Grid Engine File Formats SGE_CONF(5)
2
3
4
6 sge_conf - Grid Engine bootstrap file
7
9 bootstrap contains parameters that are needed for the startup of Grid
10 Engine components. It is created during the sge_qmaster installation.
11 Modifying bootstrap in a running system is not supported.
12
14 The paragraphs that follow provide brief descriptions of the individual
15 parameters that compose the bootstrap configuration for a Grid Engine
16 cluster:
17
18 admin_user
19 Administrative user account used by Grid Engine for all internal file
20 handling (status spooling, message logging, etc.). Can be used in cases
21 where the root account does not have the corresponding file access per‐
22 missions (e.g. on a shared file system without global root read/write
23 access).
24
25 Changing the admin_user parameter set at installation time is not sup‐
26 ported in a running system. Changing it manually on a shut-down cluster
27 is possible, but if access to the Grid Engine spooling area is inter‐
28 rupted, this will result in unpredictable behavior.
29
30 The admin_user parameter has no default value, but instead it is
31 defined during the master installation procedure.
32
33 default_domain
34 Only needed if your Grid Engine cluster covers hosts belonging to more
35 than a single DNS domain. In this case it can be used if your hostname
36 resolving yields both qualified and unqualified hostnames for the hosts
37 in one of the DNS domains. The value of default_domain is appended to
38 the unqualified hostname to define a fully qualified hostname. The
39 default_domain parameter will have no effect if ignore_fqdn is set to
40 "true".
41
42 Changing the default_domain parameter set at installation time is not
43 supported in a running system. The default for default_domain is
44 "none", in which case it will not be used.
45
46 ignore_fqdn
47 Ignore fully qualified domain name component of hostnames. Should be
48 set if all hosts belonging to a Grid Engine cluster are part of a sin‐
49 gle DNS domain. It is switched on if set to either "true" or "1".
50 Switching it on may solve problems with load reports due to different
51 hostname resolutions across the cluster.
52
53 Changing the ignore_fqdn parameter set at installation time is not sup‐
54 ported in a running system. The default for ignore_fqdn is "true".
55
56 spooling_method
57 Defines how sge_qmaster(8) writes its configuration and the status
58 information of a running cluster.
59
60 The available spooling methods are berkeleydb and classic.
61
62 spooling_lib
63 The name of a shared library containing the spooling_method to be
64 loaded at sge_qmaster(8) initialization time. The extension character‐
65 izing a shared library (.so, .sl, .dylib etc.) is not contained in
66 spooling_lib.
67
68 If spooling_method was set to berkeleydb during installation, spool‐
69 ing_lib is set to libspoolb. If classic was chosen as spooling_method,
70 spooling_lib is set to libspoolc.
71
72 Not all operating systems allow the dynamic loading of libraries. On
73 these platforms a certain spooling method (default: berkeleydb) is com‐
74 piled into the binaries and the parameter spooling_lib will be ignored.
75
76 spooling_params
77 Defines parameters to the chosen spooling method.
78
79 Parameters that are needed to initialize the spooling framework, e.g.
80 to open database files or to connect to a certain database server.
81
82 The spooling parameters value for spooling method berkeleydb is
83 [rpc_server:]database directory, e.g. /sge_local/default/spool/qmas‐
84 ter/spooldb for spooling to a local filesystem, or myhost:sge for
85 spooling over a Berkeley DB RPC server.
86
87 For spooling method classic the spooling parameters take the form <com‐
88 mon_dir>;<qmaster spool dir>, e.g. /sge/default/com‐
89 mon;/sge/default/spool/qmaster
90
91 binary_path
92 The directory path where the Grid Engine binaries reside. It is used
93 within Grid Engine components to locate and startup other Grid Engine
94 programs.
95
96 The path name given here is searched for binaries as well as any direc‐
97 tory below with a directory name equal to the current operating system
98 architecture. Therefore, /usr/SGE/bin will work for all architectures,
99 if the corresponding binaries are located in subdirectories named
100 aix43, cray, lx24-x86, hp11, irix65, tru64, sol-sparc, etc.
101
102 The default location for the binary path is <SGE_ROOT>/bin
103
104 qmaster_spool_dir
105 The location where the master spool directory resides. Only the
106 sge_qmaster(8) and sge_shadowd(8) need to have access to this direc‐
107 tory. The master spool directory - in particular the job_scripts
108 directory and the messages log file - may become quite large depending
109 on the size of the cluster and the number of jobs. Be sure to allocate
110 enough disk space and regularly clean off the log files, e.g. via a
111 cron(8) job.
112
113 Changing the qmaster_spool_dir parameter set at installation time is
114 not supported in a running system.
115
116 The default location for the master spool directory is
117 $SGE_ROOT/$SGE_CELL/spool/qmaster.
118
119 product_mode
120 The product mode defines the set of features the installed cluster is
121 using.
122
123 Both the basic product modes sge (Grid Engine standard mode) and sgeee
124 (Grid Engine Enterprise Edition) as well as additional features (AFS,
125 DCE, KERBEROS, RESERVED_PORT security model, CSP security model) are
126 specified by this parameter.
127
129 See sge_intro(1) for a full statement of rights and permissions.
130
131
132
133GE 6.1 $Date: 2007/07/19 08:17:17 $ SGE_CONF(5)