1 HAIL(1) Ganeti | Version @GANETI_VERSION@
2 =========================================
7 hail - Ganeti IAllocator plugin
12 **hail** [ **-t** *file* | **--simulate** *spec* ] [options...] *input-file*
19 hail is a Ganeti IAllocator plugin that allows automatic instance
20 placement and automatic instance secondary node replacement using the
21 same algorithm as **hbal**(1).
23 The program takes input via a JSON-file containing current cluster
24 state and the request details, and output (on stdout) a JSON-formatted
25 response. In case of critical failures, the error message is printed
26 on stderr and the exit code is changed to show failure.
31 The program uses a simplified version of the hbal algorithm.
33 For relocations, we try to change the secondary node of the instance
34 to all the valid other nodes; the node which results in the best
35 cluster score is chosen.
37 For single-node allocations (non-mirrored instances), again we
38 select the node which, when chosen as the primary node, gives the best
41 For dual-node allocations (mirrored instances), we chose the best
42 pair; this is the only choice where the algorithm is non-trivial
43 with regard to cluster size.
45 For node evacuations (*multi-evacuate* mode), we iterate over all
46 instances which live as secondaries on those nodes and try to relocate
47 them using the single-instance relocation algorithm.
49 In all cases, the cluster scoring is identical to the hbal algorithm.
54 The options that can be passed to the program are as follows:
57 Prints the before and after node status, in a format designed to
58 allow the user to understand the node's most important
59 parameters. See the man page **hbal**(1) for more details about this
62 -t *datafile*, --text-data=*datafile*
63 The name of the file holding cluster information, to override the
64 data in the JSON request itself. This is mostly used for debugging.
66 --simulate *description*
67 Similar to the **-t** option, this allows overriding the cluster
68 data with a simulated cluster. For details about the description,
69 see the man page **hspace**(1).
71 -S *filename*, --save-cluster=*filename*
72 If given, the state of the cluster before and the iallocator run is
73 saved to a file named *filename.pre-ialloc*, respectively
74 *filename.post-ialloc*. This allows re-feeding the cluster state to
75 any of the htools utilities.
78 This option increases verbosity and can be used for debugging in order
79 to understand how the IAllocator request is parsed; it can be passed
80 multiple times for successively more information.
86 For the tag-exclusion configuration (see the manpage of hbal for more
87 details), the list of which instance tags to consider as exclusion
88 tags will be read from the cluster tags, configured as follows:
90 - get all cluster tags starting with **htools:iextags:**
91 - use their suffix as the prefix for exclusion tags
93 For example, given a cluster tag like **htools:iextags:service**,
94 all instance tags of the form **service:X** will be considered as
95 exclusion tags, meaning that (e.g.) two instances which both have a
96 tag **service:foo** will not be placed on the same primary node.
101 The options that can be passed to the program are as follows:
106 The exist status of the command will be zero, unless for some reason
107 the algorithm fatally failed (e.g. wrong node or instance data).
109 .. vim: set textwidth=72 :