4 # Copyright (C) 2006, 2007 Google Inc.
6 # This program is free software; you can redistribute it and/or modify
7 # it under the terms of the GNU General Public License as published by
8 # the Free Software Foundation; either version 2 of the License, or
9 # (at your option) any later version.
11 # This program is distributed in the hope that it will be useful, but
12 # WITHOUT ANY WARRANTY; without even the implied warranty of
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 # General Public License for more details.
16 # You should have received a copy of the GNU General Public License
17 # along with this program; if not, write to the Free Software
18 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
22 # pylint: disable-msg=W0401,W0614
23 # W0401: Wildcard import ganeti.cli
24 # W0614: Unused import %s from wildcard import (since we need cli)
30 from ganeti.cli import *
31 from ganeti import opcodes
32 from ganeti import constants
33 from ganeti import errors
34 from ganeti import utils
35 from ganeti import bootstrap
36 from ganeti import ssh
37 from ganeti import objects
41 def InitCluster(opts, args):
42 """Initialize the cluster.
44 @param opts: the command line options selected by the user
46 @param args: should contain only one element, the desired
49 @return: the desired exit code
52 if not opts.lvm_storage and opts.vg_name:
53 ToStderr("Options --no-lvm-storage and --vg-name conflict.")
56 vg_name = opts.vg_name
57 if opts.lvm_storage and not opts.vg_name:
58 vg_name = constants.DEFAULT_VG
60 hvlist = opts.enabled_hypervisors
62 hvlist = constants.DEFAULT_ENABLED_HYPERVISOR
63 hvlist = hvlist.split(",")
65 hvparams = dict(opts.hvparams)
66 beparams = opts.beparams
67 nicparams = opts.nicparams
69 # prepare beparams dict
70 beparams = objects.FillDict(constants.BEC_DEFAULTS, beparams)
71 utils.ForceDictType(beparams, constants.BES_PARAMETER_TYPES)
73 # prepare nicparams dict
74 nicparams = objects.FillDict(constants.NICC_DEFAULTS, nicparams)
75 utils.ForceDictType(nicparams, constants.NICS_PARAMETER_TYPES)
77 # prepare hvparams dict
78 for hv in constants.HYPER_TYPES:
79 if hv not in hvparams:
81 hvparams[hv] = objects.FillDict(constants.HVC_DEFAULTS[hv], hvparams[hv])
82 utils.ForceDictType(hvparams[hv], constants.HVS_PARAMETER_TYPES)
84 if opts.candidate_pool_size is None:
85 opts.candidate_pool_size = constants.MASTER_POOL_SIZE_DEFAULT
87 if opts.mac_prefix is None:
88 opts.mac_prefix = constants.DEFAULT_MAC_PREFIX
90 bootstrap.InitCluster(cluster_name=args[0],
91 secondary_ip=opts.secondary_ip,
93 mac_prefix=opts.mac_prefix,
94 master_netdev=opts.master_netdev,
95 file_storage_dir=opts.file_storage_dir,
96 enabled_hypervisors=hvlist,
100 candidate_pool_size=opts.candidate_pool_size,
101 modify_etc_hosts=opts.modify_etc_hosts,
103 op = opcodes.OpPostInitCluster()
109 def DestroyCluster(opts, args):
110 """Destroy the cluster.
112 @param opts: the command line options selected by the user
114 @param args: should be an empty list
116 @return: the desired exit code
119 if not opts.yes_do_it:
120 ToStderr("Destroying a cluster is irreversible. If you really want"
121 " destroy this cluster, supply the --yes-do-it option.")
124 op = opcodes.OpDestroyCluster()
125 master = SubmitOpCode(op)
126 # if we reached this, the opcode didn't fail; we can proceed to
127 # shutdown all the daemons
128 bootstrap.FinalizeClusterDestroy(master)
132 def RenameCluster(opts, args):
133 """Rename the cluster.
135 @param opts: the command line options selected by the user
137 @param args: should contain only one element, the new cluster name
139 @return: the desired exit code
144 usertext = ("This will rename the cluster to '%s'. If you are connected"
145 " over the network to the cluster name, the operation is very"
146 " dangerous as the IP address will be removed from the node"
147 " and the change may not go through. Continue?") % name
148 if not AskUser(usertext):
151 op = opcodes.OpRenameCluster(name=name)
156 def RedistributeConfig(opts, args):
157 """Forces push of the cluster configuration.
159 @param opts: the command line options selected by the user
161 @param args: empty list
163 @return: the desired exit code
166 op = opcodes.OpRedistributeConfig()
167 SubmitOrSend(op, opts)
171 def ShowClusterVersion(opts, args):
172 """Write version of ganeti software to the standard output.
174 @param opts: the command line options selected by the user
176 @param args: should be an empty list
178 @return: the desired exit code
182 result = cl.QueryClusterInfo()
183 ToStdout("Software version: %s", result["software_version"])
184 ToStdout("Internode protocol: %s", result["protocol_version"])
185 ToStdout("Configuration format: %s", result["config_version"])
186 ToStdout("OS api version: %s", result["os_api_version"])
187 ToStdout("Export interface: %s", result["export_version"])
191 def ShowClusterMaster(opts, args):
192 """Write name of master node to the standard output.
194 @param opts: the command line options selected by the user
196 @param args: should be an empty list
198 @return: the desired exit code
201 master = bootstrap.GetMaster()
205 def _PrintGroupedParams(paramsdict):
206 """Print Grouped parameters (be, nic, disk) by group.
208 @type paramsdict: dict of dicts
209 @param paramsdict: {group: {param: value, ...}, ...}
212 for gr_name, gr_dict in paramsdict.items():
213 ToStdout(" - %s:", gr_name)
214 for item, val in gr_dict.iteritems():
215 ToStdout(" %s: %s", item, val)
217 def ShowClusterConfig(opts, args):
218 """Shows cluster information.
220 @param opts: the command line options selected by the user
222 @param args: should be an empty list
224 @return: the desired exit code
228 result = cl.QueryClusterInfo()
230 ToStdout("Cluster name: %s", result["name"])
232 ToStdout("Creation time: %s", utils.FormatTime(result["ctime"]))
233 ToStdout("Modification time: %s", utils.FormatTime(result["mtime"]))
235 ToStdout("Master node: %s", result["master"])
237 ToStdout("Architecture (this node): %s (%s)",
238 result["architecture"][0], result["architecture"][1])
241 tags = ", ".join(utils.NiceSort(result["tags"]))
245 ToStdout("Tags: %s", tags)
247 ToStdout("Default hypervisor: %s", result["default_hypervisor"])
248 ToStdout("Enabled hypervisors: %s", ", ".join(result["enabled_hypervisors"]))
250 ToStdout("Hypervisor parameters:")
251 _PrintGroupedParams(result["hvparams"])
253 ToStdout("Cluster parameters:")
254 ToStdout(" - candidate pool size: %s", result["candidate_pool_size"])
255 ToStdout(" - master netdev: %s", result["master_netdev"])
256 ToStdout(" - lvm volume group: %s", result["volume_group_name"])
257 ToStdout(" - file storage path: %s", result["file_storage_dir"])
259 ToStdout("Default instance parameters:")
260 _PrintGroupedParams(result["beparams"])
262 ToStdout("Default nic parameters:")
263 _PrintGroupedParams(result["nicparams"])
268 def ClusterCopyFile(opts, args):
269 """Copy a file from master to some nodes.
271 @param opts: the command line options selected by the user
273 @param args: should contain only one element, the path of
274 the file to be copied
276 @return: the desired exit code
280 if not os.path.exists(filename):
281 raise errors.OpPrereqError("No such filename '%s'" % filename)
285 myname = utils.HostInfo().name
287 cluster_name = cl.QueryConfigValues(["cluster_name"])[0]
289 results = GetOnlineNodes(nodes=opts.nodes, cl=cl)
290 results = [name for name in results if name != myname]
292 srun = ssh.SshRunner(cluster_name=cluster_name)
294 if not srun.CopyFileToNode(node, filename):
295 ToStderr("Copy of file %s to node %s failed", filename, node)
300 def RunClusterCommand(opts, args):
301 """Run a command on some nodes.
303 @param opts: the command line options selected by the user
305 @param args: should contain the command to be run and its arguments
307 @return: the desired exit code
312 command = " ".join(args)
314 nodes = GetOnlineNodes(nodes=opts.nodes, cl=cl)
316 cluster_name, master_node = cl.QueryConfigValues(["cluster_name",
319 srun = ssh.SshRunner(cluster_name=cluster_name)
321 # Make sure master node is at list end
322 if master_node in nodes:
323 nodes.remove(master_node)
324 nodes.append(master_node)
327 result = srun.Run(name, "root", command)
328 ToStdout("------------------------------------------------")
329 ToStdout("node: %s", name)
330 ToStdout("%s", result.output)
331 ToStdout("return code = %s", result.exit_code)
336 def VerifyCluster(opts, args):
337 """Verify integrity of cluster, performing various test on nodes.
339 @param opts: the command line options selected by the user
341 @param args: should be an empty list
343 @return: the desired exit code
347 if opts.skip_nplusone_mem:
348 skip_checks.append(constants.VERIFY_NPLUSONE_MEM)
349 op = opcodes.OpVerifyCluster(skip_checks=skip_checks,
350 verbose=opts.verbose,
351 error_codes=opts.error_codes,
352 debug_simulate_errors=opts.simulate_errors)
359 def VerifyDisks(opts, args):
360 """Verify integrity of cluster disks.
362 @param opts: the command line options selected by the user
364 @param args: should be an empty list
366 @return: the desired exit code
369 op = opcodes.OpVerifyDisks()
370 result = SubmitOpCode(op)
371 if not isinstance(result, (list, tuple)) or len(result) != 3:
372 raise errors.ProgrammerError("Unknown result type for OpVerifyDisks")
374 bad_nodes, instances, missing = result
376 retcode = constants.EXIT_SUCCESS
379 for node, text in bad_nodes.items():
380 ToStdout("Error gathering data on node %s: %s",
381 node, utils.SafeEncode(text[-400:]))
383 ToStdout("You need to fix these nodes first before fixing instances")
386 for iname in instances:
389 op = opcodes.OpActivateInstanceDisks(instance_name=iname)
391 ToStdout("Activating disks for instance '%s'", iname)
393 except errors.GenericError, err:
394 nret, msg = FormatError(err)
396 ToStderr("Error activating disks for instance %s: %s", iname, msg)
399 for iname, ival in missing.iteritems():
400 all_missing = utils.all(ival, lambda x: x[0] in bad_nodes)
402 ToStdout("Instance %s cannot be verified as it lives on"
403 " broken nodes", iname)
405 ToStdout("Instance %s has missing logical volumes:", iname)
407 for node, vol in ival:
408 if node in bad_nodes:
409 ToStdout("\tbroken node %s /dev/xenvg/%s", node, vol)
411 ToStdout("\t%s /dev/xenvg/%s", node, vol)
412 ToStdout("You need to run replace_disks for all the above"
413 " instances, if this message persist after fixing nodes.")
419 def RepairDiskSizes(opts, args):
420 """Verify sizes of cluster disks.
422 @param opts: the command line options selected by the user
424 @param args: optional list of instances to restrict check to
426 @return: the desired exit code
429 op = opcodes.OpRepairDiskSizes(instances=args)
434 def MasterFailover(opts, args):
435 """Failover the master node.
437 This command, when run on a non-master node, will cause the current
438 master to cease being master, and the non-master to become new
441 @param opts: the command line options selected by the user
443 @param args: should be an empty list
445 @return: the desired exit code
449 usertext = ("This will perform the failover even if most other nodes"
450 " are down, or if this node is outdated. This is dangerous"
451 " as it can lead to a non-consistent cluster. Check the"
452 " gnt-cluster(8) man page before proceeding. Continue?")
453 if not AskUser(usertext):
456 return bootstrap.MasterFailover(no_voting=opts.no_voting)
459 def SearchTags(opts, args):
460 """Searches the tags on all the cluster.
462 @param opts: the command line options selected by the user
464 @param args: should contain only one element, the tag pattern
466 @return: the desired exit code
469 op = opcodes.OpSearchTags(pattern=args[0])
470 result = SubmitOpCode(op)
473 result = list(result)
475 for path, tag in result:
476 ToStdout("%s %s", path, tag)
479 def SetClusterParams(opts, args):
480 """Modify the cluster.
482 @param opts: the command line options selected by the user
484 @param args: should be an empty list
486 @return: the desired exit code
489 if not (not opts.lvm_storage or opts.vg_name or
490 opts.enabled_hypervisors or opts.hvparams or
491 opts.beparams or opts.nicparams or
492 opts.candidate_pool_size is not None):
493 ToStderr("Please give at least one of the parameters.")
496 vg_name = opts.vg_name
497 if not opts.lvm_storage and opts.vg_name:
498 ToStdout("Options --no-lvm-storage and --vg-name conflict.")
500 elif not opts.lvm_storage:
503 hvlist = opts.enabled_hypervisors
504 if hvlist is not None:
505 hvlist = hvlist.split(",")
507 # a list of (name, dict) we can pass directly to dict() (or [])
508 hvparams = dict(opts.hvparams)
509 for hv, hv_params in hvparams.iteritems():
510 utils.ForceDictType(hv_params, constants.HVS_PARAMETER_TYPES)
512 beparams = opts.beparams
513 utils.ForceDictType(beparams, constants.BES_PARAMETER_TYPES)
515 nicparams = opts.nicparams
516 utils.ForceDictType(nicparams, constants.NICS_PARAMETER_TYPES)
518 op = opcodes.OpSetClusterParams(vg_name=vg_name,
519 enabled_hypervisors=hvlist,
523 candidate_pool_size=opts.candidate_pool_size)
528 def QueueOps(opts, args):
531 @param opts: the command line options selected by the user
533 @param args: should contain only one element, the subcommand
535 @return: the desired exit code
540 if command in ("drain", "undrain"):
541 drain_flag = command == "drain"
542 client.SetQueueDrainFlag(drain_flag)
543 elif command == "info":
544 result = client.QueryConfigValues(["drain_flag"])
549 ToStdout("The drain flag is %s" % val)
551 raise errors.OpPrereqError("Command '%s' is not valid." % command)
556 def _ShowWatcherPause(until):
557 if until is None or until < time.time():
558 ToStdout("The watcher is not paused.")
560 ToStdout("The watcher is paused until %s.", time.ctime(until))
563 def WatcherOps(opts, args):
564 """Watcher operations.
566 @param opts: the command line options selected by the user
568 @param args: should contain only one element, the subcommand
570 @return: the desired exit code
576 if command == "continue":
577 client.SetWatcherPause(None)
578 ToStdout("The watcher is no longer paused.")
580 elif command == "pause":
582 raise errors.OpPrereqError("Missing pause duration")
584 result = client.SetWatcherPause(time.time() + ParseTimespec(args[1]))
585 _ShowWatcherPause(result)
587 elif command == "info":
588 result = client.QueryConfigValues(["watcher_pause"])
589 _ShowWatcherPause(result)
592 raise errors.OpPrereqError("Command '%s' is not valid." % command)
599 InitCluster, [ArgHost(min=1, max=1)],
600 [DEBUG_OPT, BACKEND_OPT, CP_SIZE_OPT, ENABLED_HV_OPT, GLOBAL_FILEDIR_OPT,
601 HVLIST_OPT, MAC_PREFIX_OPT, MASTER_NETDEV_OPT, NIC_PARAMS_OPT,
602 NOLVM_STORAGE_OPT, NOMODIFY_ETCHOSTS_OPT, SECONDARY_IP_OPT, VG_NAME_OPT],
603 "[opts...] <cluster_name>", "Initialises a new cluster configuration"),
605 DestroyCluster, ARGS_NONE,
606 [DEBUG_OPT, YES_DOIT_OPT],
607 "", "Destroy cluster"),
609 RenameCluster, [ArgHost(min=1, max=1)],
610 [DEBUG_OPT, FORCE_OPT],
612 "Renames the cluster"),
614 RedistributeConfig, ARGS_NONE, [DEBUG_OPT, SUBMIT_OPT],
615 "", "Forces a push of the configuration file and ssconf files"
616 " to the nodes in the cluster"),
618 VerifyCluster, ARGS_NONE,
619 [DEBUG_OPT, VERBOSE_OPT, DEBUG_SIMERR_OPT, ERROR_CODES_OPT, NONPLUS1_OPT],
620 "", "Does a check on the cluster configuration"),
622 VerifyDisks, ARGS_NONE, [DEBUG_OPT],
623 "", "Does a check on the cluster disk status"),
624 'repair-disk-sizes': (
625 RepairDiskSizes, ARGS_MANY_INSTANCES, [DEBUG_OPT],
626 "", "Updates mismatches in recorded disk sizes"),
628 MasterFailover, ARGS_NONE, [DEBUG_OPT, NOVOTING_OPT],
629 "", "Makes the current node the master"),
631 ShowClusterVersion, ARGS_NONE, [DEBUG_OPT],
632 "", "Shows the cluster version"),
634 ShowClusterMaster, ARGS_NONE, [DEBUG_OPT],
635 "", "Shows the cluster master"),
637 ClusterCopyFile, [ArgFile(min=1, max=1)],
638 [DEBUG_OPT, NODE_LIST_OPT],
639 "[-n node...] <filename>", "Copies a file to all (or only some) nodes"),
641 RunClusterCommand, [ArgCommand(min=1)],
642 [DEBUG_OPT, NODE_LIST_OPT],
643 "[-n node...] <command>", "Runs a command on all (or only some) nodes"),
645 ShowClusterConfig, ARGS_NONE, [DEBUG_OPT],
646 "", "Show cluster configuration"),
649 [DEBUG_OPT], "", "List the tags of the cluster"),
651 AddTags, [ArgUnknown()], [DEBUG_OPT, TAG_SRC_OPT],
652 "tag...", "Add tags to the cluster"),
654 RemoveTags, [ArgUnknown()], [DEBUG_OPT, TAG_SRC_OPT],
655 "tag...", "Remove tags from the cluster"),
657 SearchTags, [ArgUnknown(min=1, max=1)],
658 [DEBUG_OPT], "", "Searches the tags on all objects on"
659 " the cluster for a given pattern (regex)"),
662 [ArgChoice(min=1, max=1, choices=["drain", "undrain", "info"])],
663 [DEBUG_OPT], "drain|undrain|info", "Change queue properties"),
666 [ArgChoice(min=1, max=1, choices=["pause", "continue", "info"]),
667 ArgSuggest(min=0, max=1, choices=["30m", "1h", "4h"])],
669 "{pause <timespec>|continue|info}", "Change watcher properties"),
671 SetClusterParams, ARGS_NONE,
672 [DEBUG_OPT, BACKEND_OPT, CP_SIZE_OPT, ENABLED_HV_OPT, HVLIST_OPT,
673 NIC_PARAMS_OPT, NOLVM_STORAGE_OPT, VG_NAME_OPT],
675 "Alters the parameters of the cluster"),
678 if __name__ == '__main__':
679 sys.exit(GenericMain(commands, override={"tag_type": constants.TAG_CLUSTER}))