Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 61d67afe

History | View | Annotate | Download (5.3 kB)

1
Version 0.1.3 (Fri, 05 Jun 2009):
2
  - Fix a bug in the ReplacePrimary instance moves, affecting most of
3
    the tools
4

    
5
Version 0.1.2 (Tue, 02 Jun 2009):
6
  - Add a new program, “hspace”, which computes the free space on a
7
    cluster (based on a given instance spec)
8
  - Improvements in API docs and partially in the user docs
9
  - Started adding unittests
10

    
11
Version 0.1.1 (Tue, 26 May 2009):
12
  - Add a new program, “hail”, which is an iallocator plugin and can
13
    allocate/relocate instances
14
  - Experimental support for non-mirrored instances (hail supports them,
15
    hbal should no longer abort when it finds such instances and simply
16
    ignore them)
17
  - The RAPI port and/or scheme can be overriden now, and even “file://”
18
    schemes can be used if the message body has been saved under the
19
    appropriate name
20
  - Lots of code reorganization, esp. rewritten loading pipeline
21
  - Better data checking and better error messages in case validation
22
    fails; tools now consider nodes with error in input data (‘?’
23
    returned by ganeti) as offline
24
  - Small enhancement to the makefile for simpler packaging
25

    
26
Version 0.1.0 (Tue, 19 May 2009):
27
  - Drop compatibility with Ganeti 1.2
28
  - Add a new minimum score option (with a very low default), should
29
    help with very good clusters (but is still not optimal)
30
  - Add a --quiet option to hbal
31
  - Add support for reading offline nodes directly from the cluster
32

    
33
Version 0.0.8 (Tue, 21 Apr 2009):
34
  - hbal: prevent mismatches in wrong node names being passed to -O, by
35
    aborting in this case
36
  - add the ability to write the commands (-C) to a script via
37
    (-C<file>), so that it can be later executed directly; this has also
38
    changed the commands to include the ncessary -f flags to skip
39
    confirmations
40
  - add checks for extra argument in hbal and hn1, so that unintended
41
    errors are catched
42
  - raise the accepted “missing” memory limit to 512MB, to cover usual
43
    Xen reservations
44

    
45
Version 0.0.7 (Mon, 23 Mar 2009):
46
  - added support for offline nodes, which are not used as targets for
47
    instance relocation and if they hold instances the hbal algorithm
48
    will attempt to relocate these away
49
  - added support for offline instances, which now will no longer skew
50
    the free memory estimation of nodes; the algorithm will no longer
51
    create conditions for N+1 failures when such instances are later
52
    started
53
  - implemented a complete model of node resources, in order to
54
    prevent an unintended re-occurrence of cases like the offline
55
    instance were we miscalculate some node resource; this gives warning
56
    now in case the node reported free disk or free memory deviates by
57
    more than a set amount from the expected value
58
  - a new tool *hscan* that can generate the input text-file for the
59
    other tools by collection via RAPI
60
  - some small changes to the build system to make it more friendly;
61
    also included the generated documentation in the source archive
62

    
63
Version 0.0.6 (Mon, 16 Mar 2009):
64
  - re-factored the hbal algorithm to make it stable in the sense that it
65
    gives the same solution when restarted from the middle; barring
66
    rounding of disk/memory and incomplete reporting from Ganeti (for
67
    1.2), it should be now feasible to rely on its output without
68
    generating moves ad infinitum
69
  - the hbal algorithm now uses two more variables: the node N+1
70
    failures and the amount of reserved memory; the first of which tries
71
    to ‘fix’ the N+1 status, the latter tries to distribute secondaries
72
    more equally
73
  - the hbal algorithm now uses two more moves at each step:
74
    replace+failover and failover+replace (besides the original
75
    failover, replace, and failover+replace+failover)
76
  - slightly changed the build system to embed GIT version/tags into the
77
    binaries so that we know for a binary from which tree it was done,
78
    either via ‘--version’ or via “strings hbal|grep version”
79
  - changed the solution list and in general the hbal output to be more
80
    clear by default, and changed “gnt-instance failover” to
81
    “gnt-instance migrate”
82
  - added man pages for the two binaries
83

    
84
Version 0.0.5 (Mon, 09 Mar 2009):
85
  - a few small improvements for hbal (possibly undone by later
86
    changes), hbal is now quite faster
87
  - fix documentation building
88
  - allow hbal to work on non N+1 compliant clusters, but without
89
    guarantees that the end cluster will be compliant; in any case, this
90
    should give a smaller number of nodes that are not compliant if the
91
    cluster state permits it
92
  - strip common domain suffix from nodes and instances, so that output
93
    is shorter and hopefully clearer
94

    
95
Version 0.0.4 (Sun, 15 Feb 2009):
96
  - better balancing algorithm in hbal
97
  - implemented an RAPI collector, now the cluster data can be gathered
98
    automatically via RAPI and doesn't need manual export of node and
99
    instance list
100

    
101
Version 0.0.3 (Wed, 28 Jan 2009):
102
  - initial release of the hbal, a cluster rebalancing tool
103
  - input data format changed due to hbal requirements
104

    
105
Version 0.0.2 (Tue, 06 Jan 2009):
106
  - fix handling of some common cases (cluster N+1 compliant from the
107
    start, too big depth given, failure to compute solution)
108
  - add option to print the needed command list for reaching the
109
    proposed solution
110

    
111
Version 0.0.1 (Tue, 06 Jan 2009):
112
  - initial release of hn1 tool