1 .TH HBAL 1 2009-03-14 htools "Ganeti H-tools"
3 hbal \- Cluster balancer for Ganeti
11 .BI "[ -m " cluster "]"
12 .BI "[-n " nodes-file " ]"
13 .BI "[ -i " instances-file "]"
19 hbal is a cluster balancer that looks at the current state of the
20 cluster (nodes with their total and free disk, memory, etc.) and
21 instance placement and computes a series of steps designed to bring
22 the cluster into a better state.
24 The algorithm to do so is designed to be stable (i.e. it will give you
25 the same results when restarting it from the middle of the solution)
26 and reasonably fast. It is not, however, designed to be a perfect
27 algorithm - it is possible to make it go into a corner from which it
28 can find no improvement, because it only look one "step" ahead.
30 By default, the program will show the solution incrementally as it is
31 computed, in a somewhat cryptic format; for getting the actual Ganeti
32 command list, use the \fB-C\fR option.
36 The program works in independent steps; at each step, we compute the
37 best instance move that lowers the cluster score.
39 The possible move type for an instance are combinations of
40 failover/migrate and replace-disks such that we change one of the
41 instance nodes, and the other one remains (but possibly with changed
42 role, e.g. from primary it becomes secondary). The list is:
44 - replace secondary (r)
45 - replace primary, a composite move (f, r, f)
46 - failover and replace secondary, also composite (f, r)
47 - replace secondary and failover, also composite (r, f)
49 We don't do the only remaining possibility of replacing both nodes
50 (r,f,r,f or the equivalent f,r,f,r) since these move needs an
51 exhaustive search over both candidate primary and secondary nodes, and
52 is O(n*n) in the number of nodes. Furthermore, it doesn't seems to
53 give better scores but will result in more disk replacements.
57 As said before, the algorithm tries to minimise the cluster score at
58 each step. Currently this score is computed as a sum of the following
60 - coefficient of variance of the percent of free memory
61 - coefficient of variance of the percent of reserved memory
62 - coefficient of variance of the percent of free disk
63 - percentage of nodes failing N+1 check
65 The free memory and free disk values help ensure that all nodes are
66 somewhat balanced in their resource usage. The reserved memory helps
67 to ensure that nodes are somewhat balanced in holding secondary
68 instances, and that no node keeps too much memory reserved for
69 N+1. And finally, the N+1 percentage helps guide the algorithm towards
70 eliminating N+1 failures, if possible.
72 Except for the N+1 failures, we use the coefficient of variance since
73 this brings the values into the same unit so to speak, and with a
74 restrict domain of values (between zero and one). The percentage of
75 N+1 failures, while also in this numeric range, doesn't actually has
76 the same meaning, but it has shown to work well.
78 The other alternative, using for N+1 checks the coefficient of
79 variance of (N+1 fail=1, N+1 pass=0) across nodes could hint the
80 algorithm to make more N+1 failures if most nodes are N+1 fail
81 already. Since this (making N+1 failures) is not allowed by other
82 rules of the algorithm, so the N+1 checks would simply not work
85 On a perfectly balanced cluster (all nodes the same size, all
86 instances the same size and spread across the nodes equally), all
87 values would be zero. This doesn't happen too often in practice :)
89 .SS OTHER POSSIBLE METRICS
91 It would be desirable to add more metrics to the algorithm, especially
92 dynamically-computed metrics, such as:
93 - CPU usage of instances, combined with VCPU versus PCPU count
98 The options that can be passed to the program are as follows:
100 .B -C, --print-commands
101 Print the command list at the end of the run. Without this, the
102 program will only show a shorter, but cryptic output.
105 Prints the before and after node status, in a format designed to allow
106 the user to understand the node's most important parameters.
108 The node list will contain these informations:
109 - a character denoting the N+1 status of the node, with blank
110 meaning pass and an asterisk ('*') meaning fail
112 - the total node memory
113 - the free node memory
114 - the reserved node memory, which is the amount of free memory
115 needed for N+1 compliance
118 - number of primary instances
119 - number of secondary instances
120 - percent of free memory
121 - percent of free disk
125 Only shows a one-line output from the program, designed for the case
126 when one wants to look at multiple clusters at once and check their
129 The line will contain four fields:
130 - initial cluster score
131 - number of steps in the solution
132 - final cluster score
133 - improvement in the cluster score
136 .BI "-n" nodefile ", --nodes=" nodefile
137 The name of the file holding node information (if not collecting via
138 RAPI), instead of the default
143 .BI "-i" instancefile ", --instances=" instancefile
144 The name of the file holding instance information (if not collecting
145 via RAPI), instead of the default
151 Collect data not from files but directly from the
153 given as an argument via RAPI. This work for both Ganeti 1.2 and
157 .BI "-l" N ", --max-length=" N
158 Restrict the solution to this length. This can be used for example to
159 automate the execution of the balancing.
163 Increase the output verbosity. Each usage of this option will increase
164 the verbosity (currently more than 2 doesn't make sense) from the
169 Just show the program version and exit.
173 The exist status of the command will be zero, unless for some reason
174 the algorithm fatally failed (e.g. wrong node or instance data).
178 The program does not check its input data for consistency, and aborts
179 with cryptic errors messages in this case.
181 The algorithm is not perfect.
183 The output format is not easily scriptable, and the program should
184 feed moves directly into Ganeti (either via RAPI or via a gnt-debug
191 With the default options, the program shows each individual step and
192 the improvements it brings in cluster score:
197 Loaded 20 nodes, 80 instances
198 Cluster is not N+1 happy, continuing but no guarantee that the cluster will end N+1 happy.
199 Initial score: 0.52329131
200 Trying to minimize the CV...
201 1. instance14 node1:node10 => node16:node10 0.42109120 a=f r:node16 f
202 2. instance54 node4:node15 => node16:node15 0.31904594 a=f r:node16 f
203 3. instance4 node5:node2 => node2:node16 0.26611015 a=f r:node16
204 4. instance48 node18:node20 => node2:node18 0.21361717 a=r:node2 f
205 5. instance93 node19:node18 => node16:node19 0.16166425 a=r:node16 f
206 6. instance89 node3:node20 => node2:node3 0.11005629 a=r:node2 f
207 7. instance5 node6:node2 => node16:node6 0.05841589 a=r:node16 f
208 8. instance94 node7:node20 => node20:node16 0.00658759 a=f r:node16
209 9. instance44 node20:node2 => node2:node15 0.00438740 a=f r:node15
210 10. instance62 node14:node18 => node14:node16 0.00390087 a=r:node16
211 11. instance13 node11:node14 => node11:node16 0.00361787 a=r:node16
212 12. instance19 node10:node11 => node10:node7 0.00336636 a=r:node7
213 13. instance43 node12:node13 => node12:node1 0.00305681 a=r:node1
214 14. instance1 node1:node2 => node1:node4 0.00263124 a=r:node4
215 15. instance58 node19:node20 => node19:node17 0.00252594 a=r:node17
216 Cluster score improved from 0.52329131 to 0.00252594
220 In the above output, we can see:
221 - the input data (here from files) shows a cluster with 20 nodes and
223 - the cluster is not initially N+1 compliant
224 - the initial score is 0.52329131
226 The step list follows, showing the instance, its initial
227 primary/secondary nodes, the new primary secondary, the cluster list,
228 and the actions taken in this step (with 'f' denoting failover/migrate
229 and 'r' denoting replace secondary).
231 Finally, the program shows the improvement in cluster score.
233 A more detailed output is obtained via the \fB-C\fR and \fB-p\fR options:
238 Loaded 20 nodes, 80 instances
239 Cluster is not N+1 happy, continuing but no guarantee that the cluster will end N+1 happy.
240 Initial cluster status:
241 N1 Name t_mem f_mem r_mem t_dsk f_dsk pri sec p_fmem p_fdsk
242 * node1 32762 1280 6000 1861 1026 5 3 0.03907 0.55179
243 node2 32762 31280 12000 1861 1026 0 8 0.95476 0.55179
244 * node3 32762 1280 6000 1861 1026 5 3 0.03907 0.55179
245 * node4 32762 1280 6000 1861 1026 5 3 0.03907 0.55179
246 * node5 32762 1280 6000 1861 978 5 5 0.03907 0.52573
247 * node6 32762 1280 6000 1861 1026 5 3 0.03907 0.55179
248 * node7 32762 1280 6000 1861 1026 5 3 0.03907 0.55179
249 node8 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
250 node9 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
251 * node10 32762 7280 12000 1861 1026 4 4 0.22221 0.55179
252 node11 32762 7280 6000 1861 922 4 5 0.22221 0.49577
253 node12 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
254 node13 32762 7280 6000 1861 922 4 5 0.22221 0.49577
255 node14 32762 7280 6000 1861 922 4 5 0.22221 0.49577
256 * node15 32762 7280 12000 1861 1131 4 3 0.22221 0.60782
257 node16 32762 31280 0 1861 1860 0 0 0.95476 1.00000
258 node17 32762 7280 6000 1861 1106 5 3 0.22221 0.59479
259 * node18 32762 1280 6000 1396 561 5 3 0.03907 0.40239
260 * node19 32762 1280 6000 1861 1026 5 3 0.03907 0.55179
261 node20 32762 13280 12000 1861 689 3 9 0.40535 0.37068
263 Initial score: 0.52329131
264 Trying to minimize the CV...
265 1. instance14 node1:node10 => node16:node10 0.42109120 a=f r:node16 f
266 2. instance54 node4:node15 => node16:node15 0.31904594 a=f r:node16 f
267 3. instance4 node5:node2 => node2:node16 0.26611015 a=f r:node16
268 4. instance48 node18:node20 => node2:node18 0.21361717 a=r:node2 f
269 5. instance93 node19:node18 => node16:node19 0.16166425 a=r:node16 f
270 6. instance89 node3:node20 => node2:node3 0.11005629 a=r:node2 f
271 7. instance5 node6:node2 => node16:node6 0.05841589 a=r:node16 f
272 8. instance94 node7:node20 => node20:node16 0.00658759 a=f r:node16
273 9. instance44 node20:node2 => node2:node15 0.00438740 a=f r:node15
274 10. instance62 node14:node18 => node14:node16 0.00390087 a=r:node16
275 11. instance13 node11:node14 => node11:node16 0.00361787 a=r:node16
276 12. instance19 node10:node11 => node10:node7 0.00336636 a=r:node7
277 13. instance43 node12:node13 => node12:node1 0.00305681 a=r:node1
278 14. instance1 node1:node2 => node1:node4 0.00263124 a=r:node4
279 15. instance58 node19:node20 => node19:node17 0.00252594 a=r:node17
280 Cluster score improved from 0.52329131 to 0.00252594
282 Commands to run to reach the above solution:
284 echo gnt-instance migrate instance14
285 echo gnt-instance replace-disks -n node16 instance14
286 echo gnt-instance migrate instance14
288 echo gnt-instance migrate instance54
289 echo gnt-instance replace-disks -n node16 instance54
290 echo gnt-instance migrate instance54
292 echo gnt-instance migrate instance4
293 echo gnt-instance replace-disks -n node16 instance4
295 echo gnt-instance replace-disks -n node2 instance48
296 echo gnt-instance migrate instance48
298 echo gnt-instance replace-disks -n node16 instance93
299 echo gnt-instance migrate instance93
301 echo gnt-instance replace-disks -n node2 instance89
302 echo gnt-instance migrate instance89
304 echo gnt-instance replace-disks -n node16 instance5
305 echo gnt-instance migrate instance5
307 echo gnt-instance migrate instance94
308 echo gnt-instance replace-disks -n node16 instance94
310 echo gnt-instance migrate instance44
311 echo gnt-instance replace-disks -n node15 instance44
313 echo gnt-instance replace-disks -n node16 instance62
315 echo gnt-instance replace-disks -n node16 instance13
317 echo gnt-instance replace-disks -n node7 instance19
319 echo gnt-instance replace-disks -n node1 instance43
321 echo gnt-instance replace-disks -n node4 instance1
323 echo gnt-instance replace-disks -n node17 instance58
325 Final cluster status:
326 N1 Name t_mem f_mem r_mem t_dsk f_dsk pri sec p_fmem p_fdsk
327 node1 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
328 node2 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
329 node3 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
330 node4 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
331 node5 32762 7280 6000 1861 1078 4 5 0.22221 0.57947
332 node6 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
333 node7 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
334 node8 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
335 node9 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
336 node10 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
337 node11 32762 7280 6000 1861 1022 4 4 0.22221 0.54951
338 node12 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
339 node13 32762 7280 6000 1861 1022 4 4 0.22221 0.54951
340 node14 32762 7280 6000 1861 1022 4 4 0.22221 0.54951
341 node15 32762 7280 6000 1861 1031 4 4 0.22221 0.55408
342 node16 32762 7280 6000 1861 1060 4 4 0.22221 0.57007
343 node17 32762 7280 6000 1861 1006 5 4 0.22221 0.54105
344 node18 32762 7280 6000 1396 761 4 2 0.22221 0.54570
345 node19 32762 7280 6000 1861 1026 4 4 0.22221 0.55179
346 node20 32762 13280 6000 1861 1089 3 5 0.40535 0.58565
351 Here we see, beside the step list, the initial and final cluster
352 status, with the final one showing all nodes being N+1 compliant, and
353 the command list to reach the final solution. In the initial listing,
354 we see which nodes are not N+1 compliant.
356 The algorithm is stable as long as each step above is fully completed,
357 e.g. in step 8, both the migrate and the replace-disks are
358 done. Otherwise, if only the migrate is done, the input data is
359 changed in a way that the program will output a different solution
360 list (but hopefully will end in the same state).
363 hn1(1), ganeti(7), gnt-instance(8), gnt-node(8)