1 Version 0.1.4 (Tue, 16 Jun 2009):
2 - Added CPU count/ratio of virtual-to-physical CPUs to the cluster
3 scoring methods; this means that now the balancer, the iallocator
4 plugin and so on will try to keep the VCPU-to-PCPU ratio equal
6 - Fixed some hscan bugs
7 - Fixed the way iallocator reads the total disk size (was broken and
8 it was always falling back to summing the disk sizes)
9 - Internals: fixed most compile-time warnings
11 Version 0.1.3 (Fri, 05 Jun 2009):
12 - Fix a bug in the ReplacePrimary instance moves, affecting most of
15 Version 0.1.2 (Tue, 02 Jun 2009):
16 - Add a new program, “hspace”, which computes the free space on a
17 cluster (based on a given instance spec)
18 - Improvements in API docs and partially in the user docs
19 - Started adding unittests
21 Version 0.1.1 (Tue, 26 May 2009):
22 - Add a new program, “hail”, which is an iallocator plugin and can
23 allocate/relocate instances
24 - Experimental support for non-mirrored instances (hail supports them,
25 hbal should no longer abort when it finds such instances and simply
27 - The RAPI port and/or scheme can be overriden now, and even “file://”
28 schemes can be used if the message body has been saved under the
30 - Lots of code reorganization, esp. rewritten loading pipeline
31 - Better data checking and better error messages in case validation
32 fails; tools now consider nodes with error in input data (‘?’
33 returned by ganeti) as offline
34 - Small enhancement to the makefile for simpler packaging
36 Version 0.1.0 (Tue, 19 May 2009):
37 - Drop compatibility with Ganeti 1.2
38 - Add a new minimum score option (with a very low default), should
39 help with very good clusters (but is still not optimal)
40 - Add a --quiet option to hbal
41 - Add support for reading offline nodes directly from the cluster
43 Version 0.0.8 (Tue, 21 Apr 2009):
44 - hbal: prevent mismatches in wrong node names being passed to -O, by
46 - add the ability to write the commands (-C) to a script via
47 (-C<file>), so that it can be later executed directly; this has also
48 changed the commands to include the ncessary -f flags to skip
50 - add checks for extra argument in hbal and hn1, so that unintended
52 - raise the accepted “missing” memory limit to 512MB, to cover usual
55 Version 0.0.7 (Mon, 23 Mar 2009):
56 - added support for offline nodes, which are not used as targets for
57 instance relocation and if they hold instances the hbal algorithm
58 will attempt to relocate these away
59 - added support for offline instances, which now will no longer skew
60 the free memory estimation of nodes; the algorithm will no longer
61 create conditions for N+1 failures when such instances are later
63 - implemented a complete model of node resources, in order to
64 prevent an unintended re-occurrence of cases like the offline
65 instance were we miscalculate some node resource; this gives warning
66 now in case the node reported free disk or free memory deviates by
67 more than a set amount from the expected value
68 - a new tool *hscan* that can generate the input text-file for the
69 other tools by collection via RAPI
70 - some small changes to the build system to make it more friendly;
71 also included the generated documentation in the source archive
73 Version 0.0.6 (Mon, 16 Mar 2009):
74 - re-factored the hbal algorithm to make it stable in the sense that it
75 gives the same solution when restarted from the middle; barring
76 rounding of disk/memory and incomplete reporting from Ganeti (for
77 1.2), it should be now feasible to rely on its output without
78 generating moves ad infinitum
79 - the hbal algorithm now uses two more variables: the node N+1
80 failures and the amount of reserved memory; the first of which tries
81 to ‘fix’ the N+1 status, the latter tries to distribute secondaries
83 - the hbal algorithm now uses two more moves at each step:
84 replace+failover and failover+replace (besides the original
85 failover, replace, and failover+replace+failover)
86 - slightly changed the build system to embed GIT version/tags into the
87 binaries so that we know for a binary from which tree it was done,
88 either via ‘--version’ or via “strings hbal|grep version”
89 - changed the solution list and in general the hbal output to be more
90 clear by default, and changed “gnt-instance failover” to
91 “gnt-instance migrate”
92 - added man pages for the two binaries
94 Version 0.0.5 (Mon, 09 Mar 2009):
95 - a few small improvements for hbal (possibly undone by later
96 changes), hbal is now quite faster
97 - fix documentation building
98 - allow hbal to work on non N+1 compliant clusters, but without
99 guarantees that the end cluster will be compliant; in any case, this
100 should give a smaller number of nodes that are not compliant if the
101 cluster state permits it
102 - strip common domain suffix from nodes and instances, so that output
103 is shorter and hopefully clearer
105 Version 0.0.4 (Sun, 15 Feb 2009):
106 - better balancing algorithm in hbal
107 - implemented an RAPI collector, now the cluster data can be gathered
108 automatically via RAPI and doesn't need manual export of node and
111 Version 0.0.3 (Wed, 28 Jan 2009):
112 - initial release of the hbal, a cluster rebalancing tool
113 - input data format changed due to hbal requirements
115 Version 0.0.2 (Tue, 06 Jan 2009):
116 - fix handling of some common cases (cluster N+1 compliant from the
117 start, too big depth given, failure to compute solution)
118 - add option to print the needed command list for reaching the
121 Version 0.0.1 (Tue, 06 Jan 2009):
122 - initial release of hn1 tool
projects / ganeti-local / blob