Advertisement
Guest User

Opensm /h

a guest
Mar 4th, 2014
85
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 12.88 KB | None | 0 0
  1. -------------------------------------------------
  2. OpenSM 3.3.13 UMAD
  3. Command Line Arguments:
  4.  
  5. ------- OpenSM - Usage and options ----------------------
  6. Usage: opensm [options]
  7. Options:
  8. --version
  9. Prints OpenSM version and exits.
  10.  
  11. --config, -F <file-name>
  12. The name of the OpenSM config file. When not specified
  13. '%ProgramFiles%\OFED\OpenSM\opensm.conf' will be used (if exists).
  14.  
  15. --create-config, -c <file-name>
  16. OpenSM will dump its configuration to the specified file and exit.
  17. This is a way to generate OpenSM configuration file template.
  18.  
  19. --guid, -g <GUID in hex>
  20. This option specifies the local port GUID value
  21. with which OpenSM should bind. OpenSM may be
  22. bound to 1 port at a time.
  23. If GUID given is 0, OpenSM displays a list
  24. of possible port GUIDs and waits for user input.
  25. Without -g, OpenSM tries to use the default port.
  26.  
  27. --lmc, -l <LMC>
  28. This option specifies the subnet's LMC value.
  29. The number of LIDs assigned to each port is 2^LMC.
  30. The LMC value must be in the range 0-7.
  31. LMC values > 0 allow multiple paths between ports.
  32. LMC values > 0 should only be used if the subnet
  33. topology actually provides multiple paths between
  34. ports, i.e. multiple interconnects between switches.
  35. Without -l, OpenSM defaults to LMC = 0, which allows
  36. one path between any two ports.
  37.  
  38. --priority, -p <PRIORITY>
  39. This option specifies the SM's PRIORITY.
  40. This will effect the handover cases, where master
  41. is chosen by priority and GUID. Range goes
  42. from 0 (lowest priority) to 15 (highest).
  43.  
  44. --smkey, -k <SM_Key>
  45. This option specifies the SM's SM_Key (64 bits).
  46. This will effect SM authentication.
  47. Note that OpenSM version 3.2.1 and below used the
  48. default value '1' in a host byte order, it is fixed
  49. now but you may need this option to interoperate
  50. with old OpenSM running on a little endian machine.
  51.  
  52. --reassign_lids, -r
  53. This option causes OpenSM to reassign LIDs to all
  54. end nodes. Specifying -r on a running subnet
  55. may disrupt subnet traffic.
  56. Without -r, OpenSM attempts to preserve existing
  57. LID assignments resolving multiple use of same LID.
  58.  
  59. --routing_engine, -R <engine name>
  60. This option chooses routing engine(s) to use instead of default
  61. Min Hop algorithm. Multiple routing engines can be specified
  62. separated by commas so that specific ordering of routing
  63. algorithms will be tried if earlier routing engines fail.
  64. If all configured routing engines fail, OpenSM will always
  65. attempt to route with Min Hop unless 'no_fallback' is
  66. included in the list of routing engines.
  67. Supported engines: updn, dnup, file, ftree, lash, dor, torus-2QoS
  68.  
  69. --do_mesh_analysis
  70. This option enables additional analysis for the lash
  71. routing engine to precondition switch port assignments
  72. in regular cartesian meshes which may reduce the number
  73. of SLs required to give a deadlock free routing
  74.  
  75. --lash_start_vl <vl number>
  76. Sets the starting VL to use for the lash routing algorithm.
  77. Defaults to 0.
  78. --sm_sl <sl number>
  79. Sets the SL to use to communicate with the SM/SA. Defaults to 0.
  80.  
  81. --connect_roots, -z
  82. This option enforces routing engines (up/down and
  83. fat-tree) to make connectivity between root switches
  84. and in this way be IBA compliant. In many cases,
  85. this can violate "pure" deadlock free algorithm, so
  86. use it carefully.
  87.  
  88. --ucast_cache, -A
  89. This option enables unicast routing cache to prevent
  90. routing recalculation (which is a heavy task in a
  91. large cluster) when there was no topology change
  92. detected during the heavy sweep, or when the topology
  93. change does not require new routing calculation,
  94. e.g. in case of host reboot.
  95. This option becomes very handy when the cluster size
  96. is thousands of nodes.
  97.  
  98. --lid_matrix_file, -M <file name>
  99. This option specifies the name of the lid matrix dump file
  100. from where switch lid matrices (min hops tables will be
  101. loaded.
  102.  
  103. --lfts_file, -U <file name>
  104. This option specifies the name of the LFTs file
  105. from where switch forwarding tables will be loaded.
  106.  
  107. --sadb_file, -S <file name>
  108. This option specifies the name of the SA DB dump file
  109. from where SA database will be loaded.
  110.  
  111. --root_guid_file, -a <path to file>
  112. Set the root nodes for the Up/Down or Fat-Tree routing
  113. algorithm to the guids provided in the given file (one
  114. to a line)
  115.  
  116. --cn_guid_file, -u <path to file>
  117. Set the compute nodes for the Fat-Tree routing algorithm
  118. to the guids provided in the given file (one to a line)
  119.  
  120. --io_guid_file, -G <path to file>
  121. Set the I/O nodes for the Fat-Tree routing algorithm
  122. to the guids provided in the given file (one to a line)
  123.  
  124. --port-shifting
  125. Attempt to shift port routes around to remove alignment problems
  126. in routing tables
  127.  
  128. --scatter-ports <random seed>
  129. Randomize best port chosen for a route
  130.  
  131. --max_reverse_hops, -H <hop_count>
  132. Set the max number of hops the wrong way around
  133. an I/O node is allowed to do (connectivity for I/O nodes on top swithces)
  134.  
  135. --ids_guid_file, -m <path to file>
  136. Name of the map file with set of the IDs which will be used
  137. by Up/Down routing algorithm instead of node GUIDs
  138. (format: <guid> <id> per line)
  139.  
  140. --guid_routing_order_file, -X <path to file>
  141. Set the order port guids will be routed for the MinHop
  142. and Up/Down routing algorithms to the guids provided in the
  143. given file (one to a line)
  144.  
  145. --torus_config <path to file>
  146. This option defines the file name for the extra configuration
  147. info needed for the torus-2QoS routing engine. The default
  148. name is '%ProgramFiles%\OFED\OpenSM\osm-torus-2QoS.conf'
  149.  
  150. --once, -o
  151. This option causes OpenSM to configure the subnet
  152. once, then exit. Ports remain in the ACTIVE state.
  153.  
  154. --sweep, -s <interval>
  155. This option specifies the number of seconds between
  156. subnet sweeps. Specifying -s 0 disables sweeping.
  157. Without -s, OpenSM defaults to a sweep interval of
  158. 10 seconds.
  159.  
  160. --timeout, -t <milliseconds>
  161. This option specifies the time in milliseconds
  162. used for transaction timeouts.
  163. Timeout values should be > 0.
  164. Without -t, OpenSM defaults to a timeout value of
  165. 200 milliseconds.
  166.  
  167. --retries <number>
  168. This option specifies the number of retries used
  169. for transactions.
  170. Without --retries, OpenSM defaults to 3 retries
  171. for transactions.
  172.  
  173. --maxsmps, -n <number>
  174. This option specifies the number of VL15 SMP MADs
  175. allowed on the wire at any one time.
  176. Specifying --maxsmps 0 allows unlimited outstanding
  177. SMPs.
  178. Without --maxsmps, OpenSM defaults to a maximum of
  179. 4 outstanding SMPs.
  180.  
  181. --console, -q [off|local]
  182. This option activates the OpenSM console (default off).
  183.  
  184. --ignore-guids, -i <equalize-ignore-guids-file>
  185. This option provides the means to define a set of ports
  186. (by guid) that will be ignored by the link load
  187. equalization algorithm.
  188.  
  189. --hop_weights_file, -w <path to file>
  190. This option provides the means to define a weighting
  191. factor per port for customizing the least weight
  192. hops for the routing.
  193.  
  194. --port_search_ordering_file, -O <path to file>
  195. This option provides the means to define a mapping
  196. between ports and dimension (Order) for controlling
  197. Dimension Order Routing (DOR).
  198. Moreover this option provides the means to define non
  199. default routing port order.
  200.  
  201. --dimn_ports_file, -O <path to file> (DEPRECATED)
  202. Use --port_search_ordering_file instead.
  203. This option provides the means to define a mapping
  204. between ports and dimension (Order) for controlling
  205. Dimension Order Routing (DOR).
  206.  
  207. --honor_guid2lid, -x
  208. This option forces OpenSM to honor the guid2lid file,
  209. when it comes out of Standby state, if such file exists
  210. under OSM_CACHE_DIR, and is valid. By default, this is FALSE.
  211.  
  212. --log_file, -f <log-file-name>
  213. This option defines the log to be the given file.
  214. By default, the log goes to '%windir%\temp\osm.log'.
  215. For the log to go to standard output use -f stdout.
  216.  
  217. --log_limit, -L <size in MB>
  218. This option defines maximal log file size in MB. When
  219. specified the log file will be truncated upon reaching
  220. this limit.
  221.  
  222. --erase_log_file, -e
  223. This option will cause deletion of the log file
  224. (if it previously exists). By default, the log file
  225. is accumulative.
  226.  
  227. --Pconfig, -P <partition-config-file>
  228. This option defines the optional partition configuration file.
  229. The default name is 'ProgramFiles\OFED\OpenSM\partitions.conf'.
  230.  
  231. --no_part_enforce, -N
  232. This option disables partition enforcement on switch external ports.
  233.  
  234. --qos, -Q
  235. This option enables QoS setup.
  236.  
  237. --qos_policy_file, -Y <QoS-policy-file>
  238. This option defines the optional QoS policy file.
  239. The default name is '%ProgramFiles%\OFED\OpenSM\qos-policy.conf'.
  240.  
  241. --stay_on_fatal, -y
  242. This option will cause SM not to exit on fatal initialization
  243. issues: if SM discovers duplicated guids or 12x link with
  244. lane reversal badly configured.
  245. By default, the SM will exit on these errors.
  246.  
  247. --daemon, -B
  248. Run in daemon mode - OpenSM will run in the background.
  249.  
  250. --inactive, -I
  251. Start SM in inactive rather than normal init SM state.
  252.  
  253. --prefix_routes_file <path to file>
  254. This option specifies the prefix routes file.
  255. Prefix routes control how the SA responds to path record
  256. queries for off-subnet DGIDs. Default file is:
  257. ProgramFiles\OFED\OpenSM\prefix-routes.conf
  258.  
  259. --consolidate_ipv6_snm_req
  260. Use shared MLID for IPv6 Solicited Node Multicast groups
  261. per MGID scope and P_Key.
  262.  
  263. --log_prefix <prefix text>
  264. Prefix to syslog messages from OpenSM.
  265.  
  266. --verbose, -v
  267. This option increases the log verbosity level.
  268. The -v option may be specified multiple times
  269. to further increase the verbosity level.
  270. See the -D option for more information about
  271. log verbosity.
  272.  
  273. --V, -V
  274. This option sets the maximum verbosity level and
  275. forces log flushing.
  276. The -V is equivalent to '-D 0xFF -d 2'.
  277. See the -D option for more information about
  278. log verbosity.
  279.  
  280. --D, -D <flags>
  281. This option sets the log verbosity level.
  282. A flags field must follow the -D option.
  283. A bit set/clear in the flags enables/disables a
  284. specific log level as follows:
  285. BIT LOG LEVEL ENABLED
  286. ---- -----------------
  287. 0x01 - ERROR (error messages)
  288. 0x02 - INFO (basic messages, low volume)
  289. 0x04 - VERBOSE (interesting stuff, moderate volume)
  290. 0x08 - DEBUG (diagnostic, high volume)
  291. 0x10 - FUNCS (function entry/exit, very high volume)
  292. 0x20 - FRAMES (dumps all SMP and GMP frames)
  293. 0x40 - ROUTING (dump FDB routing information)
  294. 0x80 - currently unused.
  295. Without -D, OpenSM defaults to ERROR + INFO (0x3).
  296. Specifying -D 0 disables all messages.
  297. Specifying -D 0xFF enables all messages (see -V).
  298. High verbosity levels may require increasing
  299. the transaction timeout with the -t option.
  300.  
  301. --debug, -d <number>
  302. This option specifies a debug option.
  303. These options are not normally needed.
  304. The number following -d selects the debug
  305. option to enable as follows:
  306. OPT Description
  307. --- -----------------
  308. -d0 - Ignore other SM nodes
  309. -d1 - Force single threaded dispatching
  310. -d2 - Force log flushing after each log message
  311. -d3 - Disable multicast support
  312. -d10 - Put OpenSM in testability mode
  313. Without -d, no debug options are enabled
  314.  
  315. --help, -h, -?
  316. Display this usage info then exit.
  317.  
  318.  
  319. Extended OpenSM Service Control:
  320. 'sc control OpenSM 128' Clear OSM Log file.
  321. 'sc control OpenSM 129' Start a Heavy Sweep.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement