-
Notifications
You must be signed in to change notification settings - Fork 4
/
NEWS
156 lines (141 loc) · 7.41 KB
/
NEWS
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
## New in 3.2.1 ##
== Internals ==
!! Changes in configuration files !!
- It's now possible to define new alias DEFAULT_CONNECTOR in external section
with default_connector item.
It can be used with the following fields: taktuk connector and all remotes operations.
- It's now possible to specify the sleep time before ping in "wait reboot" step.
This is a new parameter in cluster configuration called "sleep_time_before_ping" in "boot" section.
## New in 3.2 ##
== User interface ==
* Environments files in old description format are no longer supported
A migration script is available in addons/kaenv_migration
(see doc sec/4.2.4, Environment description)
* It's now possible to export files to the server a secured way
(see doc sec/4.2.2)
* Environment's demolishing tag related commands now refers to destructive tag
== REST API ==
* It's now possible to interact with the server using a REST API (see the Network API documentation)
== Deployement environment generation scripts ==
* The ssh key addons/ssh/id_deploy is not copied on deployment kernel anymore,
it's recommanded to use the ssh agent for ssh communications
== Internals ==
* Client/Server interactions are now performed using a REST API instead of DRb
(see doc sec/5)
* The kadeploy service now run an HTTP server: ruby/Webrick (see doc sec/5.1.4)
* kadeploy3d launcher has be rewritten in ruby, new options available
(see doc sec/4.2.1)
* Clients has been rewritten in single-file ruby scripts
* The debian packaging has been reworked to fit with the standards
* It's now possible to use the Kascade software to send the environment
(see doc sec/2.4 and addons/kascade/README)
!! Changes in configuration files !!
* General configuration (doc sec/2.1)
- The general configuration file is now named server.conf
- An authentication method for administrators has to be providden
- The server can be launched in secure mode using a custom certificate/key
- Clients can be forced to export files in secure mode
- You can now specify a log file and a debug file, no more syslog
- It's now possible to specify options for tarballs manipulations
- End of deployment/reboot/power hooks are now specified in the field 'hooks'
- It's now possible to specify a threshold for the autoclean loop
- It's now possible to specify a ssh private key path to load in ssh-agent
* Clusters configuration (doc sec/2.3)
- The clusters configuration file is now named clusters.conf
* Command configuration (doc sec/2.7)
- The command configuration file is now named command.conf
* Cluster specific configuration (doc sec/2.4)
- It's now possible to specify Kascade as BroadcastEnv method
* Client configuration file (doc sec/3)
- The client configuration file is now named client.conf
- The connection to a specific server can be secure or not
!! Database update !!
* The Kadeploy3 database have to be updated with the script db/update-3.2.sql
## New in 3.1.7 ##
== User interface ==
* Environments description files are now written in YAML
A migration script is available in addons/kaenv_migration
(see doc sec/4.2.4, Environment description)
* It is now possible to deploy multi-partitioned FSArchiver images
(see doc sec/4.7)
* Kastat now have an option only to display statistics of the last deployment
* It is now possible to deploy XZ-compressed images
== Internals ==
* Partitioning operation is now performed with an administrator-providen script
(see doc sec/2.6)
* Bootloader install is now performed with an administrator-providen script
(see doc sec/2.5)
* Deployment of multi-partitioned environments (see doc sec/4.7)
* Handle of GrubPXE disks as PXE boot method (see doc sec/2.2)
* PXE profiles handling is now customizable (see doc sec/2.1,2.2)
- Each kind of reboot (local, network, user) can be made using a different
boot method (PXElinux, iPXE, GrubPXE, ...)
- The directory containing the profile is customizable
- The way to generate the filename of profiles is customizable
* Filesystems suported by the deployment environment are now explicit in the
configuration so the deployment process can be modified if the filesystem
of the deployed environment is or isn't known by the deployment environment
(see doc sec/2.4,2.4.2)
* Rewritten caching system, no more 1-hour lock on the cached files
!! Changes in configuration files !!
* General configuration (doc sec/2.1)
- The way to specify PXE configuration has changed
* Cluster specific configuration (doc sec/2.3)
- A script have to be specified for the partitioning operations
- A script have to be specified for the bootloader install
- The 'kernels' field has been moved to 'boot/kernels'
- A specific PXE header can be specified for each kind of reboot
- The string \n is no more replaced by a new line in PXE headers
(use YAML notation for multi-line headers)
- A list of supported filesystems have to be given for the deployment env.
* Clusters configuration file (doc sec/2.2)
- The fields concerning the partition maps have been removed, this
functionality is no more supported
* Changes in the partitioning maps (doc sec/2.4)
- Partition maps are not needed anymore
!! Database update !!
* The Kadeploy3 database have to be updated with the script db/update-3.1.7.sql
## New in 3.1.6 ##
== User interface ==
* The CLI display has been modified to look more clear
* Get the status of the deploying nodes by pressing the <ENTER> key
* Use variables in custom PXE profiles (see doc sec/4.2.2, Use Case 6)
* Create custom microsteps before, after or instead of an existing one
(see doc sec/4.2.2, Use Case 10)
* Deploy directly on a block device with a DD image of the entire disk
(see doc sec/4.2.2, Use Case 11)
* The demolishing_env field of environments descriptions is now a boolean
(see doc sec/4.2.4, Environment description)
* A lot of variables are now exported by Kadeploy when running scripts
(see doc sec/4.5)
== Internals ==
* New automata to handle deployments workflow:
- Specify timeouts and retries on automata's microsteps (see doc sec/2.3.2)
- Improved deployment process for DD images
- Create custom microsteps before, after or instead of an existing one
(see doc sec/2.3.2, Custom operations)
- Kexec is now available at the first reboot (SetDeploymentEnv macro-step)
(see doc sec/2.3.2, The automata macro-steps)
* Parallel commands execution:
- New TakTuk wrapper
- Execution of system commands
- Better gathering of commands statuses and outputs
* Some other minor changes:
- RC scripts are now saving launch time errors in the syslog
- Doc: Creation of (new) deployment environments (see doc sec/2.6.2)
- Doc: Do a custom partitioning (see doc sec/4.7)
- Unit test suite
- A lot of bug fixes
!! Changes in configuration files !!
* Cluster specific configuration (doc sec/2.3)
- The fields concerning the demolishing settings (threshold value, ...)
have been removed, this functionality is no more supported
- The fields 'retries' of each macro step description do now have the right
behavior, it specifies the number of retries instead of the number of tries
* Clusters configuration file (doc sec/2.2)
- For display purposes, a prefix can be associated to each cluster
* Changes in the partitioning maps (doc sec/2.4)
- You can now specify parted maps (instead of parted scripts)
- In your maps you can use several PARTTYPE<N> instead of a single one
* Pre/post install files:
- Kadeploy exports more variables when running scripts (see doc sec/4.5)