1 Ganeti-htools release notes
2 ===========================
4 Version 0.1.8 (Tue, 29 Sep 2009)
5 --------------------------------
7 - Brown-paper-bag release fixing haddock issues
9 Version 0.1.7 (Mon, 28 Sep 2009)
10 --------------------------------
12 - Fixed a bug in the Luxi backend for big responses
13 - Fixed test suite exit code in presence of test failures
14 - Changed the migrate operation to run instead failover for instances
15 which were marked as not running in the input data (this could have
16 been changed since then, but it's better than today's always migrate)
17 - Added support for 'cheap' moves only (only migrate/failover) in
19 - Added support for building without curl (thus no RAPI backend)
21 Version 0.1.6 (Wed, 19 Aug 2009)
22 --------------------------------
24 - Added support for Luxi (the native Ganeti protocol)
25 - Added support for simulated clusters (for hspace only)
26 - Added timeouts for the RAPI backend
27 - Fixed a few inconsistencies in the command line handling
28 - Fixed handling of errors while loading data
29 - The 'network' is a new dependency due to the Luxi addition
31 Version 0.1.5 (Thu, 09 Jul 2009)
32 --------------------------------
34 - Removed obsolete hn1 program; this allowed removal of a lot of
36 - Lots of changes in hspace: the output now is a shell fragment in order
37 for script to source it or parse it easier; added failure reasons;
38 optimised to use less memory for large clusters
39 - Optimized the scoring algorithm (used by all tools) so that now
40 computations should be faster
42 Version 0.1.4 (Tue, 16 Jun 2009)
43 --------------------------------
45 - Added CPU count/ratio of virtual-to-physical CPUs to the cluster
46 scoring methods; this means that now the balancer, the iallocator
47 plugin and so on will try to keep the VCPU-to-PCPU ratio equal across
49 - Fixed some hscan bugs
50 - Fixed the way iallocator reads the total disk size (was broken and it
51 was always falling back to summing the disk sizes)
52 - Internals: fixed most compile-time warnings
54 Version 0.1.3 (Fri, 05 Jun 2009)
55 --------------------------------
57 - Fix a bug in the ReplacePrimary instance moves, affecting most of the
60 Version 0.1.2 (Tue, 02 Jun 2009)
61 --------------------------------
63 - Add a new program, “hspace”, which computes the free space on a
64 cluster (based on a given instance spec)
65 - Improvements in API docs and partially in the user docs
66 - Started adding unittests
68 Version 0.1.1 (Tue, 26 May 2009)
69 --------------------------------
71 - Add a new program, “hail”, which is an iallocator plugin and can
72 allocate/relocate instances
73 - Experimental support for non-mirrored instances (hail supports them,
74 hbal should no longer abort when it finds such instances and simply
76 - The RAPI port and/or scheme can be overriden now, and even “file://”
77 schemes can be used if the message body has been saved under the
79 - Lots of code reorganization, esp. rewritten loading pipeline
80 - Better data checking and better error messages in case validation
81 fails; tools now consider nodes with error in input data (‘?’ returned
83 - Small enhancement to the makefile for simpler packaging
85 Version 0.1.0 (Tue, 19 May 2009)
86 --------------------------------
88 - Drop compatibility with Ganeti 1.2
89 - Add a new minimum score option (with a very low default), should help
90 with very good clusters (but is still not optimal)
91 - Add a --quiet option to hbal
92 - Add support for reading offline nodes directly from the cluster
94 Version 0.0.8 (Tue, 21 Apr 2009)
95 --------------------------------
97 - hbal: prevent mismatches in wrong node names being passed to -O, by
99 - add the ability to write the commands (-C) to a script via (-C<file>),
100 so that it can be later executed directly; this has also changed the
101 commands to include the ncessary -f flags to skip confirmations
102 - add checks for extra argument in hbal and hn1, so that unintended
104 - raise the accepted “missing” memory limit to 512MB, to cover usual Xen
107 Version 0.0.7 (Mon, 23 Mar 2009)
108 --------------------------------
110 - added support for offline nodes, which are not used as targets for
111 instance relocation and if they hold instances the hbal algorithm will
112 attempt to relocate these away
113 - added support for offline instances, which now will no longer skew the
114 free memory estimation of nodes; the algorithm will no longer create
115 conditions for N+1 failures when such instances are later started
116 - implemented a complete model of node resources, in order to prevent an
117 unintended re-occurrence of cases like the offline instance were we
118 miscalculate some node resource; this gives warning now in case the
119 node reported free disk or free memory deviates by more than a set
120 amount from the expected value
121 - a new tool *hscan* that can generate the input text-file for the other
122 tools by collection via RAPI
123 - some small changes to the build system to make it more friendly; also
124 included the generated documentation in the source archive
126 Version 0.0.6 (Mon, 16 Mar 2009)
127 --------------------------------
129 - re-factored the hbal algorithm to make it stable in the sense that it
130 gives the same solution when restarted from the middle; barring
131 rounding of disk/memory and incomplete reporting from Ganeti (for
132 1.2), it should be now feasible to rely on its output without
133 generating moves ad infinitum
134 - the hbal algorithm now uses two more variables: the node N+1 failures
135 and the amount of reserved memory; the first of which tries to ‘fix’
136 the N+1 status, the latter tries to distribute secondaries more
138 - the hbal algorithm now uses two more moves at each step:
139 replace+failover and failover+replace (besides the original failover,
140 replace, and failover+replace+failover)
141 - slightly changed the build system to embed GIT version/tags into the
142 binaries so that we know for a binary from which tree it was done,
143 either via ‘--version’ or via “strings hbal|grep version”
144 - changed the solution list and in general the hbal output to be more
145 clear by default, and changed “gnt-instance failover” to “gnt-instance
147 - added man pages for the two binaries
149 Version 0.0.5 (Mon, 09 Mar 2009)
150 --------------------------------
152 - a few small improvements for hbal (possibly undone by later changes),
153 hbal is now quite faster
154 - fix documentation building
155 - allow hbal to work on non N+1 compliant clusters, but without
156 guarantees that the end cluster will be compliant; in any case, this
157 should give a smaller number of nodes that are not compliant if the
158 cluster state permits it
159 - strip common domain suffix from nodes and instances, so that output is
160 shorter and hopefully clearer
162 Version 0.0.4 (Sun, 15 Feb 2009)
163 --------------------------------
165 - better balancing algorithm in hbal
166 - implemented an RAPI collector, now the cluster data can be gathered
167 automatically via RAPI and doesn't need manual export of node and
170 Version 0.0.3 (Wed, 28 Jan 2009)
171 --------------------------------
173 - initial release of the hbal, a cluster rebalancing tool
174 - input data format changed due to hbal requirements
176 Version 0.0.2 (Tue, 06 Jan 2009)
177 --------------------------------
179 - fix handling of some common cases (cluster N+1 compliant from the
180 start, too big depth given, failure to compute solution)
181 - add option to print the needed command list for reaching the proposed
184 Version 0.0.1 (Tue, 06 Jan 2009)
185 --------------------------------
187 - initial release of hn1 tool
189 .. vim: set textwidth=72 :