root / man / hroller.rst @ 4fe04580
History | View | Annotate | Download (3.7 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 |
**[ --node-tags** *tag,..* **]** |
31 |
|
32 |
Reporting options: |
33 |
|
34 |
**[ -v... | -q ]** |
35 |
**[ -S *file* ]** |
36 |
**[ --one-step-only ]** |
37 |
|
38 |
DESCRIPTION |
39 |
----------- |
40 |
|
41 |
hroller is a cluster maintenance reboot scheduler. It can calculate |
42 |
which set of nodes can be rebooted at the same time while avoiding |
43 |
having both primary and secondary nodes being rebooted at the same time. |
44 |
|
45 |
For backends that support identifying the master node (currenlty |
46 |
RAPI and LUXI), the master node is scheduled as the last node |
47 |
in the last reboot group. Apart from this restriction, larger reboot |
48 |
groups are put first. |
49 |
|
50 |
ALGORITHM FOR CALCULATING OFFLINE REBOOT GROUPS |
51 |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
52 |
|
53 |
hroller will view the nodes as vertices of an undirected graph, |
54 |
with two kind of edges. Firstly, there are edges from the primary |
55 |
to the secondary node of every instance. Secondly, two nodes are connected |
56 |
by an edge if they are the primary nodes of two instances that have the |
57 |
same secondary node. It will then color the graph using a few different |
58 |
heuristics, and return the minimum-size color set found. Node with the same |
59 |
color can then simultaneously migrate all instance off to their respective |
60 |
secondary nodes, and it is safe to reboot them simultaneously. |
61 |
|
62 |
OPTIONS |
63 |
------- |
64 |
|
65 |
For a description of the standard options check **htools**\(7) and |
66 |
**hbal**\(1). |
67 |
|
68 |
\--node-tags *tag,...* |
69 |
Restrict to nodes having at least one of the given tags. |
70 |
|
71 |
\--one-step-only |
72 |
Restrict to the first reboot group. Output the group one node per line. |
73 |
|
74 |
\--offline-maintenance |
75 |
Pretend that all instances are shutdown before the reboots are carried |
76 |
out. I.e., only edges from the primary to the secondary node of an instance |
77 |
are considered. |
78 |
|
79 |
\--force |
80 |
Do not fail, even if the master node cannot be determined. |
81 |
|
82 |
|
83 |
BUGS |
84 |
---- |
85 |
|
86 |
Offline nodes should be ignored. |
87 |
|
88 |
If instances are online the tool should refuse to do offline rolling |
89 |
maintenances, unless explicitly requested. |
90 |
|
91 |
End-to-end shelltests should be provided. |
92 |
|
93 |
Online rolling maintenances (where instance need not be shut down, but |
94 |
are migrated from node to node) are not supported yet. Hroller by design |
95 |
should support them both with and without secondary node replacement. |
96 |
|
97 |
EXAMPLES |
98 |
-------- |
99 |
|
100 |
Online Rolling reboots, using tags |
101 |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
102 |
|
103 |
Selecting by tags and getting output for one step only can be used for |
104 |
planing the next maintenance step. |
105 |
:: |
106 |
|
107 |
$ hroller --node-tags needsreboot --one-step-only -L |
108 |
'First Reboot Group' |
109 |
node1.example.com |
110 |
node3.example.com |
111 |
|
112 |
Typically these nodes would be drained and migrated. |
113 |
:: |
114 |
|
115 |
$ GROUP=`hroller --node-tags needsreboot --one-step-only --no-headers -L` |
116 |
$ for node in $GROUP; do gnt-node modify -D yes $node; done |
117 |
$ for node in $GROUP; do gnt-node migrate -f --submit $node; done |
118 |
|
119 |
After maintenance, the tags would be removed and the nodes undrained. |
120 |
|
121 |
|
122 |
Offline Rolling node reboot output |
123 |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
124 |
|
125 |
If all instances are shut down, usually larger node groups can be found. |
126 |
:: |
127 |
|
128 |
$ hroller --offline-maintainance -L |
129 |
'Node Reboot Groups' |
130 |
node1.example.com,node3.example.com,node5.example.com |
131 |
node8.example.com,node6.example.com,node2.example.com |
132 |
node7.example.com,node4.example.com |
133 |
|
134 |
.. vim: set textwidth=72 : |
135 |
.. Local Variables: |
136 |
.. mode: rst |
137 |
.. fill-column: 72 |
138 |
.. End: |