1 HTOOLS(1) Ganeti | Version @GANETI_VERSION@
2 ===========================================
7 htools - Cluster allocation and placement tools for Ganeti
19 cluster capacity computation
25 saves cluster state for later reuse
28 cluster information printer
35 ``htools`` is a suite of tools designed to help with allocation/movement
36 of instances and balancing of Ganeti clusters. ``htools`` is also the
37 generic binary that must be symlinked or hardlinked under each tool's
38 name in order to perform the different functions. Alternatively, the
39 environment variable HTOOLS can be used to set the desired role.
41 Installed as ``hbal``, it computes and optionally executes a suite of
42 instance moves in order to balance the cluster.
44 Installed as ``hcheck``, it preforms cluster checks and optionally
45 simulates rebalancing with all the ``hbal`` options available.
47 Installed as ``hspace``, it computes how many additional instances can
48 be fit on a cluster, while maintaining N+1 status. It can run on models
49 of existing clusters or of simulated clusters.
51 Installed as ``hail``, it acts as an IAllocator plugin, i.e. it is used
52 by Ganeti to compute new instance allocations and instance moves.
54 Installed as ``hscan``, it scans the local or remote cluster state and
55 saves it to files which can later be reused by the other roles.
57 Installed as ``hinfo``, it prints information about the current cluster
63 Options behave the same in all program modes, but not all program modes
64 support all options. Some common options are:
67 Prints the node status, in a format designed to allow the user to
68 understand the node's most important parameters. If the command in
69 question makes a cluster transition (e.g. balancing or allocation),
70 then usually both the initial and final node status is printed.
72 It is possible to customise the listed information by passing a
73 comma-separated list of field names to this option (the field list
74 is currently undocumented), or to extend the default field list by
75 prefixing the additional field list with a plus sign. By default,
76 the node list will contain the following information:
79 a character denoting the status of the node, with '-' meaning an
80 offline node, '*' meaning N+1 failure and blank meaning a good
90 the memory used by the node itself
93 the memory used by instances
96 amount memory which seems to be in use but cannot be determined
97 why or by which instance; usually this means that the hypervisor
98 has some overhead or that there are other reporting errors
104 the reserved node memory, which is the amount of free memory
105 needed for N+1 compliance
114 the number of physical cpus on the node
117 the number of virtual cpus allocated to primary instances
120 number of primary instances
123 number of secondary instances
126 percent of free memory
132 ratio of virtual to physical cpus
135 the dynamic CPU load (if the information is available)
138 the dynamic memory load (if the information is available)
141 the dynamic disk load (if the information is available)
144 the dynamic net load (if the information is available)
146 -t *datafile*, \--text-data=*datafile*
147 Backend specification: the name of the file holding node and instance
148 information (if not collecting via RAPI or LUXI). This or one of the
149 other backends must be selected. The option is described in the man
152 The file should contain text data, line-based, with two empty lines
153 separating sections. The lines themselves are column-based, with the
154 pipe symbol (``|``) acting as separator.
156 The first section contains group data, with two columns:
161 The second sections contains node data, with the following columns:
168 - node physical cores
169 - offline field (as ``Y`` or ``N``)
173 The third section contains instance data, with the fields:
179 - instance status (in Ganeti's format, e.g. ``running`` or ``ERROR_down``)
180 - instance ``auto_balance`` flag (see man page **gnt-instance** (7))
181 - instance primary node
182 - instance secondary node(s), if any
183 - instance disk type (e.g. ``plain`` or ``drbd``)
186 The fourth section contains the cluster tags, with one tag per line
187 (no columns/no column processing).
189 The fifth section contains the ipolicies of the cluster and the node
190 groups, in the following format (separated by ``|``):
192 - owner (empty if cluster, group name otherwise)
193 - standard, min, max instance specs, containing the following values
205 Backend specification: collect data directly from the *cluster* given
206 as an argument via RAPI. If the argument doesn't contain a colon (:),
207 then it is converted into a fully-built URL via prepending
208 ``https://`` and appending the default RAPI port, otherwise it is
209 considered a fully-specified URL and used as-is.
212 Backend specification: collect data directly from the master daemon,
213 which is to be contacted via LUXI (an internal Ganeti protocol). An
214 optional *path* argument is interpreted as the path to the unix socket
215 on which the master daemon listens; otherwise, the default path used
216 by Ganeti (configured at build time) is used.
218 -I|\--ialloc-src *path*
219 Backend specification: load data directly from an iallocator request
220 (as produced by Ganeti when doing an iallocator call). The iallocator
221 request is read from specified path.
223 \--simulate *description*
224 Backend specification: instead of using actual data, build an empty
225 cluster given a node description. The *description* parameter must be
226 a comma-separated list of five elements, describing in order:
228 - the allocation policy for this node group (*preferred*, *allocable*
229 or *unallocable*, or alternatively the short forms *p*, *a* or *u*)
230 - the number of nodes in the cluster
231 - the disk size of the nodes (default in mebibytes, units can be used)
232 - the memory size of the nodes (default in mebibytes, units can be used)
233 - the cpu core count for the nodes
234 - the spindle count for the nodes
236 An example description would be **preferred,20,100G,16g,4,2**
237 describing a 20-node cluster where each node has 100GB of disk space,
238 16GiB of memory, 4 CPU cores and 2 disk spindles. Note that all nodes
239 must have the same specs currently.
241 This option can be given multiple times, and each new use defines a
242 new node group. Hence different node groups can have different
243 allocation policies and node count/specifications.
246 Increase the output verbosity. Each usage of this option will
247 increase the verbosity (currently more than 2 doesn't make sense)
248 from the default of one.
251 Decrease the output verbosity. Each usage of this option will
252 decrease the verbosity (less than zero doesn't make sense) from the
256 Just show the program version and exit.
261 Some options accept not simply numerical values, but numerical values
262 together with a unit. By default, such unit-accepting options use
263 mebibytes. Using the lower-case letters of *m*, *g* and *t* (or their
264 longer equivalents of *mib*, *gib*, *tib*, for which case doesn't
265 matter) explicit binary units can be selected. Units in the SI system
266 can be selected using the upper-case letters of *M*, *G* and *T* (or
267 their longer equivalents of *MB*, *GB*, *TB*, for which case doesn't
270 More details about the difference between the SI and binary systems can
271 be read in the *units(7)* man page.
276 The environment variable ``HTOOLS`` can be used instead of
277 renaming/symlinking the programs; simply set it to the desired role and
278 then the name of the program is no longer used.
280 .. vim: set textwidth=72 :