Statistics
| Branch: | Tag: | Revision:

root / NEWS @ bfb03e22

History | View | Annotate | Download (5.1 kB)

1
Version 0.1.2 (Tue, 02 Jun 2009):
2
  - Add a new program, “hspace”, which computes the free space on a
3
    cluster (based on a given instance spec)
4
  - Improvements in API docs and partially in the user docs
5
  - Started adding unittests
6

    
7
Version 0.1.1 (Tue, 26 May 2009):
8
  - Add a new program, “hail”, which is an iallocator plugin and can
9
    allocate/relocate instances
10
  - Experimental support for non-mirrored instances (hail supports them,
11
    hbal should no longer abort when it finds such instances and simply
12
    ignore them)
13
  - The RAPI port and/or scheme can be overriden now, and even “file://”
14
    schemes can be used if the message body has been saved under the
15
    appropriate name
16
  - Lots of code reorganization, esp. rewritten loading pipeline
17
  - Better data checking and better error messages in case validation
18
    fails; tools now consider nodes with error in input data (‘?’
19
    returned by ganeti) as offline
20
  - Small enhancement to the makefile for simpler packaging
21

    
22
Version 0.1.0 (Tue, 19 May 2009):
23
  - Drop compatibility with Ganeti 1.2
24
  - Add a new minimum score option (with a very low default), should
25
    help with very good clusters (but is still not optimal)
26
  - Add a --quiet option to hbal
27
  - Add support for reading offline nodes directly from the cluster
28

    
29
Version 0.0.8 (Tue, 21 Apr 2009):
30
  - hbal: prevent mismatches in wrong node names being passed to -O, by
31
    aborting in this case
32
  - add the ability to write the commands (-C) to a script via
33
    (-C<file>), so that it can be later executed directly; this has also
34
    changed the commands to include the ncessary -f flags to skip
35
    confirmations
36
  - add checks for extra argument in hbal and hn1, so that unintended
37
    errors are catched
38
  - raise the accepted “missing” memory limit to 512MB, to cover usual
39
    Xen reservations
40

    
41
Version 0.0.7 (Mon, 23 Mar 2009):
42
  - added support for offline nodes, which are not used as targets for
43
    instance relocation and if they hold instances the hbal algorithm
44
    will attempt to relocate these away
45
  - added support for offline instances, which now will no longer skew
46
    the free memory estimation of nodes; the algorithm will no longer
47
    create conditions for N+1 failures when such instances are later
48
    started
49
  - implemented a complete model of node resources, in order to
50
    prevent an unintended re-occurrence of cases like the offline
51
    instance were we miscalculate some node resource; this gives warning
52
    now in case the node reported free disk or free memory deviates by
53
    more than a set amount from the expected value
54
  - a new tool *hscan* that can generate the input text-file for the
55
    other tools by collection via RAPI
56
  - some small changes to the build system to make it more friendly;
57
    also included the generated documentation in the source archive
58

    
59
Version 0.0.6 (Mon, 16 Mar 2009):
60
  - re-factored the hbal algorithm to make it stable in the sense that it
61
    gives the same solution when restarted from the middle; barring
62
    rounding of disk/memory and incomplete reporting from Ganeti (for
63
    1.2), it should be now feasible to rely on its output without
64
    generating moves ad infinitum
65
  - the hbal algorithm now uses two more variables: the node N+1
66
    failures and the amount of reserved memory; the first of which tries
67
    to ‘fix’ the N+1 status, the latter tries to distribute secondaries
68
    more equally
69
  - the hbal algorithm now uses two more moves at each step:
70
    replace+failover and failover+replace (besides the original
71
    failover, replace, and failover+replace+failover)
72
  - slightly changed the build system to embed GIT version/tags into the
73
    binaries so that we know for a binary from which tree it was done,
74
    either via ‘--version’ or via “strings hbal|grep version”
75
  - changed the solution list and in general the hbal output to be more
76
    clear by default, and changed “gnt-instance failover” to
77
    “gnt-instance migrate”
78
  - added man pages for the two binaries
79

    
80
Version 0.0.5 (Mon, 09 Mar 2009):
81
  - a few small improvements for hbal (possibly undone by later
82
    changes), hbal is now quite faster
83
  - fix documentation building
84
  - allow hbal to work on non N+1 compliant clusters, but without
85
    guarantees that the end cluster will be compliant; in any case, this
86
    should give a smaller number of nodes that are not compliant if the
87
    cluster state permits it
88
  - strip common domain suffix from nodes and instances, so that output
89
    is shorter and hopefully clearer
90

    
91
Version 0.0.4 (Sun, 15 Feb 2009):
92
  - better balancing algorithm in hbal
93
  - implemented an RAPI collector, now the cluster data can be gathered
94
    automatically via RAPI and doesn't need manual export of node and
95
    instance list
96

    
97
Version 0.0.3 (Wed, 28 Jan 2009):
98
  - initial release of the hbal, a cluster rebalancing tool
99
  - input data format changed due to hbal requirements
100

    
101
Version 0.0.2 (Tue, 06 Jan 2009):
102
  - fix handling of some common cases (cluster N+1 compliant from the
103
    start, too big depth given, failure to compute solution)
104
  - add option to print the needed command list for reaching the
105
    proposed solution
106

    
107
Version 0.0.1 (Tue, 06 Jan 2009):
108
  - initial release of hn1 tool