root / man / hroller.rst @ 52278ef9
History | View | Annotate | Download (2.4 kB)
1 |
HROLLER(1) Ganeti | Version @GANETI_VERSION@ |
---|---|
2 |
============================================ |
3 |
|
4 |
NAME |
5 |
---- |
6 |
|
7 |
hroller \- Cluster rolling maintenance scheduler for Ganeti |
8 |
|
9 |
SYNOPSIS |
10 |
-------- |
11 |
|
12 |
**hroller** {backend options...} [algorithm options...] [reporting options...] |
13 |
|
14 |
**hroller** \--version |
15 |
|
16 |
|
17 |
Backend options: |
18 |
|
19 |
{ **-m** *cluster* | **-L[** *path* **]** | **-t** *data-file* | |
20 |
**-I** *path* } |
21 |
|
22 |
**[ --force ]** |
23 |
|
24 |
Algorithm options: |
25 |
|
26 |
**[ -G *name* ]** |
27 |
|
28 |
**[ -O *name...* ]** |
29 |
|
30 |
Reporting options: |
31 |
|
32 |
**[ -v... | -q ]** |
33 |
**[ -S *file* ]** |
34 |
|
35 |
DESCRIPTION |
36 |
----------- |
37 |
|
38 |
hroller is a cluster maintenance reboot scheduler. It can calculate |
39 |
which set of nodes can be rebooted at the same time while avoiding |
40 |
having both primary and secondary nodes being rebooted at the same time. |
41 |
|
42 |
For backends that support identifying the master node (currenlty |
43 |
RAPI and LUXI), the master node is scheduled as the last node |
44 |
in the last reboot group. |
45 |
|
46 |
ALGORITHM FOR CALCULATING OFFLINE REBOOT GROUPS |
47 |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
48 |
|
49 |
hroller will view the nodes as vertices of an undirected graph, |
50 |
connecting by instances which have both a primary and a secondary node. |
51 |
It will then color the graph using a few different heuristics, and |
52 |
return the minimum-size color set found. Node with the same color don't |
53 |
share an edge, and as such don't have an instance with both primary and |
54 |
secondary node on them, so they are safe to be rebooted concurrently. |
55 |
|
56 |
OPTIONS |
57 |
------- |
58 |
|
59 |
For a description of the standard options check **htools**\(7) and |
60 |
**hbal**\(1). |
61 |
|
62 |
\--force |
63 |
Do not fail, even if the master node cannot be determined. |
64 |
|
65 |
|
66 |
BUGS |
67 |
---- |
68 |
|
69 |
Offline nodes should be ignored. |
70 |
|
71 |
If instances are online the tool should refuse to do offline rolling |
72 |
maintenances, unless explicitly requested. |
73 |
|
74 |
End-to-end shelltests should be provided. |
75 |
|
76 |
Online rolling maintenances (where instance need not be shut down, but |
77 |
are migrated from node to node) are not supported yet. Hroller by design |
78 |
should support them both with and without secondary node replacement. |
79 |
|
80 |
EXAMPLE |
81 |
------- |
82 |
|
83 |
Offline Rolling node reboot output |
84 |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
85 |
|
86 |
With the default options, the program shows one reboot group per line as |
87 |
a comma-separated list. |
88 |
:: |
89 |
|
90 |
$ hroller -L |
91 |
'Node Reboot Groups' |
92 |
node1.example.com,node3.example.com,node5.example.com |
93 |
node8.example.com,node6.example.com,node2.example.com |
94 |
node7.example.com,node4.example.com |
95 |
|
96 |
.. vim: set textwidth=72 : |
97 |
.. Local Variables: |
98 |
.. mode: rst |
99 |
.. fill-column: 72 |
100 |
.. End: |