Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 2cf61508

History | View | Annotate | Download (4.1 kB)

1
Version 0.1.0 (Tue, 19 May 2009):
2
  - Drop compatibility with Ganeti 1.2
3
  - Add a new minimum score option (with a very low default), should
4
    help with very good clusters (but is still not optimal)
5
  - Add a --quiet option to hbal
6
  - Add support for reading offline nodes directly from the cluster
7

    
8
Version 0.0.8 (Tue, 21 Apr 2009):
9
  - hbal: prevent mismatches in wrong node names being passed to -O, by
10
	aborting in this case
11
  - add the ability to write the commands (-C) to a script via (-C<file>), so
12
	that it can be later executed directly; this has also changed the commands
13
	to include the ncessary -f flags to skip confirmations
14
  - add checks for extra argument in hbal and hn1, so that unintended errors
15
	are catched
16
  - raise the accepted “missing” memory limit to 512MB, to cover usual Xen
17
	reservations
18

    
19
Version 0.0.7 (Mon, 23 Mar 2009):
20
  - added support for offline nodes, which are not used as targets for
21
    instance relocation and if they hold instances the hbal algorithm
22
    will attempt to relocate these away
23
  - added support for offline instances, which now will no longer skew
24
    the free memory estimation of nodes; the algorithm will no longer
25
    create conditions for N+1 failures when such instances are later
26
    started
27
  - implemented a complete model of node resources, in order to
28
    prevent an unintended re-occurrence of cases like the offline
29
    instance were we miscalculate some node resource; this gives warning
30
    now in case the node reported free disk or free memory deviates by
31
    more than a set amount from the expected value
32
  - a new tool *hscan* that can generate the input text-file for the
33
    other tools by collection via RAPI
34
  - some small changes to the build system to make it more friendly;
35
    also included the generated documentation in the source archive
36

    
37
Version 0.0.6 (Mon, 16 Mar 2009):
38
  - re-factored the hbal algorithm to make it stable in the sense that it
39
    gives the same solution when restarted from the middle; barring
40
    rounding of disk/memory and incomplete reporting from Ganeti (for
41
    1.2), it should be now feasible to rely on its output without
42
    generating moves ad infinitum
43
  - the hbal algorithm now uses two more variables: the node N+1
44
    failures and the amount of reserved memory; the first of which tries
45
    to ‘fix’ the N+1 status, the latter tries to distribute secondaries
46
    more equally
47
  - the hbal algorithm now uses two more moves at each step:
48
    replace+failover and failover+replace (besides the original
49
    failover, replace, and failover+replace+failover)
50
  - slightly changed the build system to embed GIT version/tags into the
51
    binaries so that we know for a binary from which tree it was done,
52
    either via ‘--version’ or via “strings hbal|grep version”
53
  - changed the solution list and in general the hbal output to be more
54
    clear by default, and changed “gnt-instance failover” to
55
    “gnt-instance migrate”
56
  - added man pages for the two binaries
57

    
58
Version 0.0.5 (Mon, 09 Mar 2009):
59
  - a few small improvements for hbal (possibly undone by later
60
    changes), hbal is now quite faster
61
  - fix documentation building
62
  - allow hbal to work on non N+1 compliant clusters, but without
63
    guarantees that the end cluster will be compliant; in any case, this
64
    should give a smaller number of nodes that are not compliant if the
65
    cluster state permits it
66
  - strip common domain suffix from nodes and instances, so that output
67
    is shorter and hopefully clearer
68

    
69
Version 0.0.4 (Sun, 15 Feb 2009):
70
  - better balancing algorithm in hbal
71
  - implemented an RAPI collector, now the cluster data can be gathered
72
    automatically via RAPI and doesn't need manual export of node and
73
    instance list
74

    
75
Version 0.0.3 (Wed, 28 Jan 2009):
76
  - initial release of the hbal, a cluster rebalancing tool
77
  - input data format changed due to hbal requirements
78

    
79
Version 0.0.2 (Tue, 06 Jan 2009):
80
  - fix handling of some common cases (cluster N+1 compliant from the
81
    start, too big depth given, failure to compute solution)
82
  - add option to print the needed command list for reaching the
83
    proposed solution
84

    
85
Version 0.0.1 (Tue, 06 Jan 2009):
86
  - initial release of hn1 tool