slapos.core 1.4.10

slapos.core
===========

The core of SlapOS.
Contains the SLAP library, and the slapgrid, slapformat, slapproxy tools.
For more information, see http://www.slapos.org.

Changes
=======

1.4.10 (2018-09-20)
-------------------
* add `--buildout-debug` command line option to `slapos node software` and
`slapos node instance` commands which starts buildout debugger on errors.
* pretty print json serialised instance parameters in `slapos proxy show`
* Add devperm plugin

1.4.9 (2018-07-31)
------------------
* slapgrid: Add tear down methods to IManager interface
* manager: Add Port Redirection manager
* proxy: create empty slaproxy database if not exits yet
* slapgrid: Add methods to SlapObject.Partition for more control on generated supervisord config

1.4.8 (2018-06-26)
------------------
* format: fix brokend parse_computer_definition
* grid.promise: kill timed out promise process if terminate is not enough
* grid.promise: avoid blocking process while sending or receiving message from queue
* grid.promise: on promise timeout fail only if the problem is occurring a second time
* slapgrid: Do not set minfds. select() does not support file descriptors greater than 1023
* slapgrid: Set the minimum number of file descriptors.

1.4.7 (2018-04-08)
------------------
* grid.promise: loadModule is now done in PromiseProcess class
* collect: fix minors bugs on collect.db and collet.reporter
* grid: fix using shutil.rmtree to delete file instead of directory
* grid: do not hide `$USER` when running buildout
* grid: do not leak file descriptors to subprocesses when running e.g. 'node software'.

1.4.6 (2018-03-29)
------------------
* grid.promise: use previous promise execution result if the promise is skipped because of periodicity.
* slapgrid: update AccessStatus of instance on Master when checking promise anomaly, if the status change.

1.4.5 (2018-03-22)
------------------
* slapos.collect.db: Create an index on user table to speed up monitor collect query.
* slapos.cli.console: support new `slapos console script.py` invocation
* slapos.grid.promise: implement a new promise design and promise launcher in slapgrid
* slapos.collect: allow connect without call boostrap, set timeout option

1.4.4 (2018-01-25)
------------------
* slap.initializeConnection: Cache master node's Hateoas URL
* slapos.grid: Declare connection_parameter_hash explicitly, UnboundLocalError may occur.
* slapos.grid: rework checkpromise method to utils so it can be reused

1.4.3 (2017-11-08)
-----------------
* slapos.cli.grid: Allow definition of different pidfiles for each software subcommand in config file
* slapos.cli.configure_local: Get template locally instead do an http request.
* slapos.cli: Update API for get person certificates and register computer
* format: fix some conflicts about tun interfaces when changing the number of partitions

1.4.2 (2017-10-02)
------------------
* slapos.collect: Make internal API usable as library for third parties

1.4.1 (2017-09-25)
------------------
* slapos.format: Introduce create_tun config option (default false)
* slapos.cli: get template directly and not reply on namespaces for register
* slapos.grid: add pluging which run instance custom script at partition pre-destroy phase

1.4.0 (2017-06-26)
------------------
* slapos.grid: Use local configuration to extend master configuration
* slapos.format: Export partition configuration for the partition
* slapos: improve logs and general cleanup
* slapos.manager: Added cpuset plugin (for cgroups)
* slapos.format: Add TUN interface support
* slapos: Implement plugin system

1.3.18 (2016-11-03)
-------------------
* update default web url of master to slapos.vifib.com

1.3.17 (2016-10-25)
-------------------
* slapos.grid: Always remove .timestamp and .slapgrid if partition is destroyed.
* slapos.proxy: Propagate parent partition state to children
* slapos.grid: Increase min space (1G)
* slapos.grid: Save slapgrid state into the partition
* slapos.format: Remove passwd call while format.
* svcbackend: explicitely call the executable instead of using Popen 'executable' keyword.
* slapos.grid: Introduce new garbage collector for instances ignored by buildout

1.3.16 (2016-09-29)
-------------------
* slapos.format: Include disk usage report. Do not divide cpu_load by number of cpu cores.
* slapos.format: set login shell for slapuser and lock login by password
* slapos.slap: Do not post same connection parameters of slaves.
* slapos.proxy: allow to update software release of partition

1.3.15 (2015-12-08)
-------------------
* slapos.collect: Include disk usage report. Do not divide cpu_load by number of cpu cores.

1.3.14 (2015-10-27)
-------------------
* slapos.grid: firewall fix bugs

1.3.13 (2015-10-26)
-------------------
* slapos.grid: firewall accpet option to specify only list of ip address/wetwork to accept and reject.

1.3.12 (2015-10-15)
-------------------
* slapos.grid: add support for firewall configuration using firewalld for partition that use tap+route interface (for kvm cluster).

1.3.11 (2015-09-25)
-------------------
* slapos.grid: support shacache-ca-file and shadir-ca-file options.

1.3.10 (2015-04-28)
-------------------

1.3.9 (2015-02-20)
------------------
* slapos.format: allow to format additional list of folder for each partition to use as data storage location.
* slapos.format: allow to create tap without bridge (when using option create_tap and tap_gateway_interface), configure ip route with generated ipv4 for tap to access guest vm from host machine.
* slapos.grid: update generated buildout file with information to acess partition data storage folder.

1.3.8 (2015-02-04)
------------------

* slapos proxy: allow to specify/override host/port from command line.

1.3.7 (2015-01-30)
------------------

* slapos.grid: Don't try to process partition if software_release_url is None. Removes noisy errors in log.
* slapos node report: retry several time when removing processes from supervisor.

1.3.6.3 (2015-01-23)
--------------------

* slapos: make forbid_supervisord_automatic_launch generic.

1.3.6.2 (2015-01-22)
--------------------

* slapos.grid.svcbackend: check if watchdog is started before restarting.

1.3.6.1 (2015-01-19)
--------------------

* slapos: allow to use supervisorctl without automatically starting supervisord.
* slapos: Create supervisor configuration when running CLI.

1.3.6 (2015-01-16)
------------------

* supervisord: allow to start with --nodaemon.
* rename : zc.buildout-bootstap.py -> zc.buildout-bootstrap.py.
* update bootstrap.py.
* slapproxy: add missing getComputerPartitionCertificate method
* slapos boot: fix error reporting when ipv6 is not available

1.3.5 (2014-12-03)
------------------

* slapos.grid: do not ALWAYS sleep for promise_timeout. Instead, poll often, and continue if promise finished. This change allows a two-folds speed improvement in processing partitions.
* slapos.format: don't chown recursively Software Releases.
* slapos.util: use find to chown in chownDirectory.

1.3.4 (2014-11-26)
------------------

* slapos.slap hateoas: get 'me' document with no cache.
* slapos.grid: report: fix unbound 'destroyed' variable.
* slapos.slap: fix __getattr__ of product collection so that product.foo works.
* slapos.cli info/list: use raw print instead of logger.

1.3.3 (2014-11-18)
------------------

* slapos.slap/slapos.proxy: Fix regression: requests library ignores empty parameters.
* slapos.proxy: fix slave support (again)

1.3.2 (2014-11-14)
------------------

* slapos.slap: parse ipv6 and adds brackets if missing. Needed for requests, that now NEEDS brackets for ipv6.
* slapos.slap: cast xml from unicode to string if it is unicode before parsing it.

1.3.1 (2014-11-13)
------------------

* slapos.proxy: fix slave support.

1.3.0 (2014-11-13)
------------------

* Introduce slapos list and slapos info CLIs.
* slapos format: fix use_unique_local_address_block feature, and put default to false in configure_local.

1.2.4.1 (2014-10-09)
--------------------

* slapos format: Don't chown partitions.
* slapos format: alter_user is true again by default.

1.2.4 (2014-09-23)
------------------

* slapos.grid: add support for retention_delay.

1.2.3.1 (2014-09-15)
--------------------

* General: Add compatibility with cliff 1.7.0.
* tests: Prevent slap tests to leak its stubs/mocks.

1.2.3 (2014-09-11)
------------------

* slapos.proxy: Add multimaster basic support.

1.2.2 (2014-09-10)
------------------

* slapos.collect: Compress historical logs and fix folder permissions.

1.2.1 (2014-08-21)
------------------

* slapproxy: add automatic migration to new database schema if needed.

1.2.0 (2014-08-18)
------------------

Note: not officially released as egg.

* slapproxy: add correct support for slaves, instance_guid, state.
* slapproxy: add getComputerPartitionStatus dummy support.
* slapproxy: add multi-nodes support

1.1.2 (2014-06-02)
------------------

* Minor fixes

1.1.1 (2014-05-23)
------------------

* Drop legacy commands
* Introduced SlapOS node Collect

1.0.5 (2014-04-29)
------------------

* Fix slapgrid commands return code
* slapos proxy start do not need to be launched as root

1.0.2.1 (2014-01-16)
--------------------

Fixes:

* Add backward compabitility in slap lib with older slapproxy (<1.0.1)

1.0.1 (2014-01-14)
------------------

New features:

* Add configure-local command for standalone slapos [Cedric de Saint Martin/Gabriel Monnerat]

Fixes:

* Fix slapproxy missing _connection_dict [Rafael Monnerat]

1.0.0 (2014-01-01)
------------------

New features:

* slapconsole: Use readline for completion and history. [Jerome Perrin]
* slapos console: support for ipython and bpython [Marco Mariani]
* Initial windows support. [Jondy Zhao]
* Support new/changed parameters in command line tools, defined in documentation. [Marco Mariani]
* Register: support for one-time authentication token. [Marco Mariani]
* New command: "slapos configure client" [Marco Mariani]
* add new "root_check" option in slapos configuration file (true by default) allowing to bypass "am I root" checks in slapos. [Cedric de Saint Martin]
* Add support for getSoftwareReleaseListFromSoftwareProduct() SLAP method. [Cedric de Saint Martin]
* Add support for Software Product in request, supply and console. [Cedric de Saint Martin]

Major Improvements:

* Major refactoring of entry points, clearly defining all possible command line parameters, separating logic from arg/conf parsing and logger setup, sanitizing most parameters, and adding help and documentation for each command. [Marco Mariani]
* Correct handling of common errors: print error message instead of traceback. [Marco Mariani]
* Dramatically speed up slapformat. [Cedric de Saint Martin]
* Remove CONFIG_SITE env var from Buildout environment, fixing support of OpenSuse 12.x. [Cedric de Saint Martin]
* RootSoftwareInstance is now the default software type. [Cedric de Saint Martin]
* Allow to use SlapOS Client for instances deployed in shared SlapOS Nodes. [Cedric de Saint Martin]

Other fixes:

* Refuse to run 'slapos node' commands as non root. [Marco Mariani]
* Register: Replace all reference to vifib by SlapOS Master. [Cedric de Saint Martin]
* Watchdog: won't call bang if bang was already called but problem has not been solved. [Cédric de Saint Martin]
* Slapgrid: avoid spurious empty lines in Popen() stdout/log. [Marco Mariani]
* Slapgrid: Properly include any partition containing any SR informations in the list of partitions to proceed. [Cedric de Saint Martin]
* Slapgrid: Remove the timestamp file after defined periodicity. Fixes odd use cases when an instance failing to process after some time is still considered as valid by the node. [Cedric de Saint Martin]
* Slapgrid: Fix scary but harmless warnings, fix grammar, remove references to ViFiB. [Cedric de Saint Martin, Jérome Perrin, Marco Mariani]
* Slapgrid: Fixes support of Python >= 2.6. [Arnaud Fontaine]
* Slapgrid: Check if SR is upload-blacklisted only if we have upload informations. [Cedric de Saint Martin]
* Slapgrid: override $HOME to be software_path or instance_path. Fix leaking files like /opt/slapgrid/.npm. [Marco Mariani]
* Slapgrid: Always retrieve certificate and key, update files if content changed. Fix "quick&dirty" manual slapos.cfg swaps (change of Node ID). [Marco Mariani]
* Slapformat: Make sure everybody can read slapos configuration directory. [Cedric de Saint Martin]
* Slapformat: Fix support of slapproxy. [Marco Mariani]
* Slapformat: slapos.xml backup: handle corrupted zip files. [Marco Mariani]
* Slapformat: Don't erase shell information for each user, every time. Allows easy debugging. [Cédric de Saint Martin]


0.35.1 (2013-02-18)
-------------------

New features:

* Add ComputerPartition._instance_guid getter in SLAP library. [Cedric de Saint Martin]
* Add ComputerPartition._instance_guid support in slapproxy. [Cedric de Saint Martin]

Fixes:

* Fix link existence check when deploying instance if SR is not correctly installed. This fixes a misleading error. [Cedric de Saint Martin]
* Improve message shown to user when requesting. [Cedric de Saint Martin]
* Raise NotReady when _requested_state doesn't exist when trying to fetch it from getter. [Cedric de Saint Martin]

0.35 (2013-02-08)
-----------------

* slapos: display version number with help. [Marco Mariani]
* slapformat: backup slapos.xml to a zip archive at every change. [Marco Mariani]
* slapformat: Don't check validity of ipv4 when trying to add address that already exists. [Cedric de Saint Martin]
* slapgrid: create and run $MD5/buildout.cfg for eaiser debugging. [Marco Mariani]
* slapgrid: keep running if cp.error() or sr.error() have issues (fixes 20130119-744D94). [Marco Mariani]
* slapgrid does not crash when there are no certificates (fixes #20130121-136C24). [Marco Mariani]
* Add slapproxy-query command. [Marco Mariani]
* Other minor typo / output fixes.

0.34 (2013-01-23)
-----------------

* networkcache: only match major release number in Debian,
fixed platform detection for Ubuntu. [Marco Mariani]
* symlink to software_release in each partition. [Marco Mariani]
* slapos client: Properly expand "~" when giving configuration file location.
[Cedric de Saint Martin]
* slapgrid: stop instances that should be stopped even if buildout and/or
reporting failed. [Cedric de Saint Martin]
* slapgrid: Don't periodically force-process a stopped instance. [Cedric de Saint Martin]
* slapgrid: Handle pid files of slapgrid launched through different entry points.
[Cedric de Saint Martin]
* Watchdog: Bang is called with correct instance certificates. [Cedric Le Ninivin]
* Watchdog: Fix watchdog call. [Cedric le Ninivin]
* Add a symlink of the used software release in each partitions. [Marco Mariani]
* slapformat is verbose by default. [Cedric de Saint Martin]
* slapproxy: Filter by instance_guid, allow computer partition renames
and change of software_type and requested_state. [Marco Mariani]
* slapproxy: Stop instance even if buildout/reporting is wrong. [Cedric de Saint Martin]
* slapproxy: implement softwareInstanceRename method. [Marco Mariani]
* slapproxy: alllow requests to software_type. [Marco Mariani]
* Many other minor fixes. See git diff for details.

0.33.1 (2012-11-05)
-------------------

* Fix "slapos console" argument parsing. [Cedric de Saint Martin]

0.33 (2012-11-02)
-----------------

* Continue to improve new entry points. The following are now functional:
- slapos node format
- slapos node start/stop/restart/tail
- slapos node supervisord/supervisorctl
- slapos node supply

and add basic usage. [Cedric de Saint Martin]
* Add support for "SLAPOS_CONFIGURATION" and SLAPOS_CLIENT_CONFIGURATION
environment variables. (commit c72a53b1) [Cédric de Saint Martin]
* --only_sr also accepts plain text URIs. [Marco Mariani]

0.32.3 (2012-10-15)
-------------------

* slapgrid: Adopt new return value strategy (0=OK, 1=failed, 2=promise failed)
(commit 5d4e1522). [Cedric de Saint Martin]
* slaplib: add requestComputer (commits 6cbe82e0, aafb86eb). [Łukasz Nowak]
* slapgrid: Add stopasgroup and killasgroup to supervisor (commit 36e0ccc0).
[Cedric de Saint Martin]
* slapproxy: don't start in debug mode by default (commit e32259c8).
[Cédric Le Ninivin
* SlapObject: ALWAYS remove tmpdir (commit a652a610). [Cedric de Saint Martin]

0.32.2 (2012-10-11)
-------------------

* slapgrid: Remove default delay, now that SlapOS Master is Fast as Light
(tm). (commit 03a85d6b8) [Cedric de Saint Martin]
* Fix watchdog entry point name, introduced in v0.31. (commit a8651ba12)
[Cedric de Saint Martin]
* slapgrid: Better filter of instances, won't process false positives anymore
(hopefully). (commit ce0a73b41) [Cedric de Saint Martin]
* Various output improvements. [Cedric de Saint Martin]

0.32.1 (2012-10-09)
-------------------

* slapgrid: Make sure error logs are sent to SlapOS master. Finish
implementation began in 0.32. [Cedric de Saint Martin]
* slapgrid: Fix Usage Report in case of not empty partition with no SR.
[Cedric de Saint Martin]

0.32 (2012-10-04)
-----------------

* Introduce new, simpler "slapos" entry point. See documentation for more
informations. Note: some functionnalities of this new entry point don't work
yet or is not as simple as it should be. [Cedric de Saint Martin, Cedric Le
Ninivin]
* Revamped "slapos request" to work like described in documentation. [Cédric
Le Ninivin, Cédric de Saint Martin]
* Rewrote slapgrid logger to always log into stdout. (commits a4d277c881,
5440626dea)[Cédric de Saint Martin]

0.31.2 (2012-10-02)
-------------------

* Update slapproxy behavior: when instance already exist, only update
partition_parameter_kw. (commit 317d5c8e0aee) [Cedric de Saint Martin]

0.31.1 (2012-10-02)
-------------------

* Fixed Watchdog call in slapgrid. [Cédric Le Ninivin]

0.31 (2012-10-02)
-------------------

* Added slapos-watchdog to bang exited and failing serices in instance
in supervisord. (commits 16b2e8b8, 1dade5cd7) [Cédric Le Ninivin]
* Add safety checks before calling SlapOS Master if mandatory instance
members of SLAP classes are not properly set. Will result in less calls to
SlapOS Master in dirty cases. (commits 5097e87c9763, 5fad6316a0f6d,
f2cd014ea8aa) [Cedric de Saint Martin]
* Add "periodicty" functionnality support for instances: if an instance has
not been processed by slapgrid after defined time, process it. (commits
7609fc7a3d, 56e1c7bfbd) [Cedric Le Ninivin]
* slapproxy: Various improvements in slave support (commits 96c6b78b67,
bcac5a397d, fbb680f53b)[Cedric Le Ninivin]
* slapgrid: bulletproof slapgrid-cp: in case one instance is bad, still
processes all other ones. (commits bac94cdb56, 77bc6c75b3d, bd68b88cc3)
[Cedric de Saint Martin]
* Add support for "upload to binary cache" URL blacklist [Cedric de Saint
Martin]
* Request on proxy are identified by requester and name (commit
0c739c3) [Cedric Le Ninivin]

0.30 (2012-09-19)
-----------------

* Add initial "slave instances" support in slapproxy. [Cedric Le Ninivin]
* slapgrid-ur fix: check for partition informations only if we have to
destroy it. [Cedric de Saint Martin]

0.29 (2012-09-18)
-----------------

* buildout: Migrate slap_connection magic instance profile part to
slap-connection, and use variables names separated with '-'. [Cedric de
Saint Martin]
* slapgrid: Add support for instance.cfg instance profiles [Cedric de Saint
Martin]
* slapgrid-ur: much less calls to master. [Cedric de Saint Martin]

0.28.9 (2012-09-18)
-------------------

* slapgrid: Don't process not updated partitions (regression introduced in
0.28.7). [Cedric de Saint Martin]

0.28.8 (2012-09-18)
-------------------

* slapgrid: Don't process free partitions (regression introduced in 0.28.7).
[Cedric de Saint Martin]

0.28.7 (2012-09-14)
-------------------

* slapgrid: --maximal_delay reappeared to be used in special cases. [Cedric
de Saint Martin]

0.28.6 (2012-09-10)
-------------------

* register now use slapos.cfg.example from master. [Cédric Le Ninivin]

0.28.5 (2012-08-23)
-------------------

* Updated slapos.cfg for register [Cédric Le Ninivin]

0.28.4 (2012-08-22)
-------------------

* Fixed egg building.

0.28.3 (2012-08-22)
-------------------

* Avoid artificial tap creation on system check. [Łukasz Nowak]

0.28.2 (2012-08-17)
-------------------

* Resolved path problem in register [Cédric Le Ninivin]


0.28.1 (2012-08-17)
-------------------

* Resolved critical naming conflict

0.28 (2012-08-17)
-----------------

* Introduce "slapos node register" command, that will register computer to
SlapOS Master (vifib.net by default) for you. [Cédric Le Ninivin]
* Set .timestamp in partitions ONLY after slapgrid thinks it's okay (promises,
...). [Cedric de Saint Martin]
* slapgrid-ur: when destroying (not reporting), only care about instances to
destroy, completely ignore others. [Cedric de Saint Martin]

0.27 (2012-08-08)
-----------------

* slapformat: Raise correct error when no IPv6 is available on selected
interface. [Cedric de Saint Martin]
* slapgrid: Introduce --only_sr and --only_cp.
- only_sr filter and force the run of a single SR, and uses url_md5
(folder_id)
- only_cp filter which computer patition, will be runned. it can be a
list, splited by comman (slappartX,slappartY ...) [Rafael Monnerat]
* slapgrid: Cleanup unused option (--usage-report-periodicity). [Cedric de
Saint Martin]
* slapgrid: --develop will work also for Computer Partitions. [Cedric de Saint
Martin]
* slaplib: setConnectionDict won't call Master if parameters haven't changed.
[Cedric de Saint Martin]

0.26.2 (2012-07-09)
-------------------

* Define UTF-8 encoding in SlapOS Node codebase, as defined in PEP-263.

0.26.1 (2012-07-06)
-------------------

* slapgrid-sr: Add --develop option to make it ignore .completed files.
* SLAP library: it is now possible to fetch whole dict of connection
parameters.
* SLAP library: it is now possible to fetch single instance parameter.
* SLAP library: change Computer and ComputerPartition behavior to have proper
caching of computer partition parameters.

0.26 (2012-07-05)
-----------------

* slapformat: no_bridge option becomes 'not create_tap'.
create_tap is true by default. So a bridge is used and tap will be created by
default. [Cedric de Saint Martin]
* Add delay for slapformat. [Cedric Le Ninivin]
* If no software_type is given, use default one (i.e fix "error 500" when
requesting new instance). [Cedric de Saint Martin]
* slapgrid: promise based software release, new api to fetch full computer
information from server. [Yingjie Xu]
* slapproxy: new api to mock full computer information [Yingjie Xu]
* slapgrid: minor fix randomise delay feature. [Yingjie Xu]
* slapgrid: optimise slapgrid-cp, run buildout only if there is an update
on server side. [Yingjie Xu]
* libslap: Allow accessing ServerError. [Vincent Pelletier]

0.25 (2012-05-16)
-----------------

* Fix support for no_bridge option in configuration files for some values:
no_bridge = false was stated as true. [Cedric de Saint Martin]
* Delay a randomized period of time before calling slapgrid. [Yingjie Xu]
* slapformat: Don't require tunctl if no_bridge is set [Leonardo Rochael]
* slapformat: remove monkey patching when creating address so that it doesn't
return false positive. [Cedric de Saint Martin]
* Various: clearer error messages.

0.24 (2012-03-29)
-----------------

* Handles different errors in a user friendly way [Cedric de Saint Martin]
* slapgrid: Supports software destruction. [Łukasz Nowak]
* slap: added support to Supply.supply state parameter (available, destroyed)
[Łukasz Nowak]

0.23 (2012-02-29)
-----------------

* slapgrid : Don't create tarball of sofwtare release when shacache is not
configured. [Yingjie Xu]

0.22 (2012-02-09)
-----------------

* slapformat : Add no-bridge feature. [Cedric de Saint Martin]
* slapgrid : Add binary cache support. [Yingjie Xu]

0.21 (2011-12-23)
-----------------

* slap: Add renaming API. [Antoine Catton]

0.20 (2011-11-24)
-----------------

* slapgrid: Support service-less parttions. [Antoine Catton]
* slapgrid: Avoid gid collision while dropping privileges. [Antoine Catton]
* slapgrid: Drop down network usage during usage reporting. [Łukasz Nowak]
* general: Add sphinx documentation. [Romain Courteaud]

0.19 (2011-11-07)
-----------------

* bang: Executable to be called by being banged computer. [Łukasz Nowak]

0.18 (2011-10-18)
-----------------

* Fix 0.17 release: missing change for slap library. [Łukasz Nowak]

0.17 (2011-10-18)
-----------------

* slap: Avoid request under the hood. [Łukasz Nowak]
* slap: ComputerPartition.bang provided. It allows to update all instances
in tree. [Łukasz Nowak]
* slap: Computer.bang provided. It allows to bang all instances on computer.
[Łukasz Nowak]

0.16 (2011-10-03)
-----------------

* slapgrid: Bugfix for slapgrid introduced in 0.15. [Łukasz Nowak]

0.15 (2011-09-27)
-----------------

* slapgrid: Sanitize environment variables as early as possible. [Arnaud
Fontaine]
* slap: Docstring bugfix. [Sebastien Robin]
* slap: Make request asynchronous call. [Łukasz Nowak]

0.14 (2011-08-31)
-----------------

* slapgrid: Implement SSL based authentication to shadir and shacache.
[Łukasz Nowak]
* slapgrid, slap: Fix usage report packing list generation. [Nicolas Godbert]

0.13 (2011-08-25)
-----------------

* slapgrid: Implement software signing and shacache upload. [Lucas Carvalho]
* slap: Support slave instances [Gabriel Monnerat]
* slapformat: Generate always address for computer [Łukasz Nowak]
* slapgrid: Support promises scripts [Antoine Catton]
* general: slapos.core gets tests. [many contributors]

0.12 (2011-07-15)
-----------------

* Include modifications that should have been included in 0.11.

0.11 (2011-07-15)
-----------------

* Bug fix : slapconsole : shorthand methods request and supply now correctly
return an object. [Cedric de Saint Martin]

0.10 (2011-07-13)
-----------------

* Fix a bug in slapconsole where request and supply shorthand methods
don't accept all needed parameters. [Cedric de Saint Martin]

0.9 (2011-07-11)
----------------

* slapconsole: Simplify usage and use configuration file. You can now
just run slapconsole and type things like "request(kvm, 'mykvm')".
[Cedric de Saint Martin]
* slapformat: Fix issue of bridge not connected with real interface on
Linux >= 2.6.39 [Arnaud Fontaine]
* slapformat: Allow to have IPv6 only interface, with bridge still supporting
local IPv4 stack. [Łukasz Nowak]

0.8 (2011-06-27)
----------------

* slapgrid: Bugfix for temporary extends cache permissions. [Łukasz Nowak]

0.7 (2011-06-27)
----------------

* slapgrid: Fallback to buildout in own search path. [Łukasz Nowak]

0.6 (2011-06-27)
----------------

* slap: Fix bug: state shall be XML encapsulated. [Łukasz Nowak]

0.5 (2011-06-24)
----------------

* slapgrid: Use temporary extends-cache directory in order to make faster
remote profile refresh. [Łukasz Nowak]

0.4 (2011-06-24)
----------------

* general: Polish requirement versions. [Arnaud Fontaine]
* general: Remove libnetworkcache. [Lucas Carvalho]
* slap: Remove not needed method from interface. [Romain Courteaud]
* slap: state parameter is accepted and transmitted to SlapOS master [Łukasz
Nowak]
* slapformat: Implement dry run. [Vincent Pelletier]
* slapgrid: Allow to select any buildout binary used to bootstrap environment.
[Łukasz Nowak]


0.3 (2011-06-14)
----------------

* slap: Implement SLA by filter_kw in OpenOrder.request. [Łukasz Nowak]
* slap: Timeout network operations. [Łukasz Nowak]
* slapformat: Make slapsoft and slapuser* system users. [Kazuhiko Shiozaki]
* slapgrid: Add more tolerance with supervisord. [Łukasz Nowak]

0.2 (2011-06-01)
----------------

* Include required files in distribution [Łukasz Nowak]

0.1 (2011-05-27)
----------------

* Merged slapos.slap, slapos.tool.console, slapos.tool.format,
slapos.tool.grid, slapos.tool.libnetworkcache and slapos.tool.proxy into one
package: slapos.core


console
-------

The slapconsole tool allows to interact with a SlapOS Master throught the SLAP
library.

For more information about SlapOS or slapconsole usages, please go to
http://community.slapos.org.

The slapconsole tool is only a bare Python console with several global variables
defined and initialized.


Initialization and configuration file
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Slapconsole allows to automatically connect to a Master using URL and SSL
certificate from given slapos.cfg.
Certificate has to be *USER* certificate, manually obtained from SlapOS master
web interface.

Slapconsole tools reads the given slapos.cfg configuration file and use the
following informations :

* Master URL is read from [slapos] in the "master_url" parameter.
* SSL Certificate is read from [slapconsole] in the "cert_file" parameter.
* SSL Key is read from [slapconsole] in the "key_file" parameter.

See slapos.cfg.example for examples.


Global functions/variables
~~~~~~~~~~~~~~~~~~~~~~~~~~

* "request()" is a shorthand for slap.registerOpenOrder().request() allowing
to request instances.
* "supply()" is a shorthand for slap.registerSupply().supply() allowing to
request software installation.

For more information about those methods, please read the SLAP library
documentation.

* "product" is an instance of slap.SoftwareProductCollection whose only goal is to retrieve
the URL of the best Software Release of a given Software Product as attribute.
for each attribute call, it will retrieve from the SlapOS Master the best
available Software Release URL and return it.

This allows to request instances in a few words, i.e::

request("http://www.url.com/path/to/current/best/known/kvm/software.cfg", "mykvm")

can be simplified into ::

request(product.kvm, "mykvm")

* "slap" is an instance of the SLAP library. It is only used for advanced usages.
"slap" instance is obtained by doing ::

slap = slapos.slap.slap()
slap.initializeConnection(config.master_url,
key_file=config.key_file, cert_file=config.cert_file)


Examples
~~~~~~~~

::

>>> # Request instance
>>> request(product.kvm, "myuniquekvm")

>>> # Request instance on specific computer
>>> request(product.kvm, "myotheruniquekvm",
filter_kw={ "computer_guid": "COMP-12345" })

>>> # Request instance, specifying parameters (here nbd_ip and nbd_port)
>>> request(product.kvm, "mythirduniquekvm",
partition_parameter_kw={"nbd_ip":"2a01:e35:2e27:460:e2cb:4eff:fed9:48dc",
"nbd_port":"1024"})

>>> # Request software installation on owned computer
>>> supply(product.kvm, "mycomputer")

>>> # Fetch existing instance status
>>> request(product.kvm, "myuniquekvm").getState()

>>> # Fetch instance information on already launched instance
>>> request(product.kvm, "myuniquekvm").getConnectionParameter("url")


Format
======

slapformat is an application to prepare SlapOS-ready node to be used inside SlapGrid Cloud.

It "formats" the machine by:

- creating users and groups
- creating bridge interface
- creating needed tap interfaces
- creating TUN interfaces
- creating needed directories with proper ownership and permissions
- (optional-manager) creating cgroup resource tree for slapos

It reads configuration from /etc/opt/slapos/slapos.cfg and formats computer
accordingly. The variables are number of partitions, IP addresses, storages
and network interfaces.

Format uploads a into configured SlapOS Master server.

Format dumps allocated resources for the partition into a JSON file per
partition ~/.slapos-resource. This file contains network interfaces,
IP address ranges and port ranges. The resource constraints can be
recursively folded.

Pluggable parts for formatting are available too. They are called Managers
and can be turned on/off via configuration property manager_list.

This program shall be only run by root.

Requirements
------------

Linux with IPv6, bridging and tap interface support.

Binaries:

* brctl
* groupadd
* ip
* useradd


grid
====

slapgrid is a client of SLAPos. SLAPos provides support for deploying a SaaS
system in a minute.
Slapgrid allows you to easily deploy instances of softwares based on buildout
profiles.
For more informations about SLAP and SLAPos, please see the SLAP documentation.


Requirements
------------

A working SLAP server with informations about your computer, in order to
retrieve them.

As Vifib servers use IPv6 only, we strongly recommend an IPv6 enabled UNIX
box.

For the same reasons, Python >= 2.6 with development headers is also strongly
recommended (IPv6 support is not complete in previous releases).

For now, gcc and glibc development headers are required to build most software
releases.


Concepts
--------

Here are the fundamental concepts of slapgrid :
A Software Release (SR) is just a software.
A Computer Partition (CP) is an instance of a Software Release.
Imagine you want to install with slapgrid some software and run it. You will
have to install the software as a Software Release, and then instantiate it,
i.e configuring it for your needs, as a Computer Partition.


How it works
------------

When run, slapgrid will authenticate to the SLAP library with a computer_id and
fetch the list of Software Releases to install or remove and Computer
Partitions to start or stop.
Then, it will process each Software Release, and each Computer Partition.
It will also periodically send to SLAP the usage report of each Computer
Partition.


Installation
------------

With easy_install::

$ easy_install slapgrid

slapgrid needs several directories to be created and configured before being
able to run : a software releases directory, and an instances directory with
configured computer partition directory(ies).
You should create for each Computer Partition directory created a specific user
and associate it with its Computer Partition directory. Each Computer Partition
directory should belongs to this specific user, with permissions of 0750.


Usage
-----

slapgrid needs several informations in order to run. You can specify them by
adding arguments to the slapgrid command line, or by putting then in a
configuration file.
Beware : you need a valid computer resource on server side.


Examples
--------

simple example :
Just run slapgrid:

$ slapgrid --instance-root /path/to/instance/root --software-root
/path/to/software_root --master-url https://some.server/some.resource
--computer-id my.computer.id


configuration file example::

[slapgrid]
instance_root = /path/to/instance/root
software_root = /path/to/software/root
master_url = https://slapos.server/slap_service
computer_id = my.computer.id

then run slapgrid::

$ slapgrid --configuration-file = path/to/configuration/file



slapos.manager
==============

Manager is a plugin-like class that is being run in multiple phases of slapos node lifecycle.

- **format**, manager can format additionally the underlaying OS
- **software**, manager can react on software installation
- **instance**, manager can update instance runtime frequently

Constructor will receive configuration of current stage. Then each method receives
object most related to the current operation. For details see <slapos/manager/interface.py>.

In code, a list of manager instances can be easily retreived by

from slapos import manager
manager_list = manager.from_config(config)

Where `from_config` extracts "manager_list" item from dict-like `config` argument
and then dynamically loads modules named according to the configuration inside
`slapos.manager` package. The manager must be a class named Manager and implementing
interface `slapos.manager.interface.IManager`.

Managers might require a list of user for whom they are allowed to perform tasks.
This list of users is given by "power_user_list" in [slapos] section in the
config file.


proxy
=====

Implement minimalist SlapOS Master server without any security, designed to work
only from localhost with one SlapOS Node (a.k.a Computer).

It implements (or should implement) the SLAP API, as currently implemented in
the SlapOS Master (see slaptool.py in Master).

The only behavioral difference from the SlapOS Master is:

When the proxy doesn't find any free partition (and/or in case of slave
instance, any compatible master instance), it will throw a NotFoundError (404).


slap
====

Simple Language for Accounting and Provisioning python library.

Developer note - python version
-------------------------------

This library is used on client (slapgrid) and server side.
Server is using python2.4 and client is using python2.6
Having this in mind, code of this library *have* to work
on python2.4

How it works
------------

The SLAP main server which is in charge of service coordination receives from participating servers the number of computer paritions which are available, the type of resource which a party is ready provide, and request from parties for resources which are needed.

Each participating server is identified by a unique ID and runs a slap-server daemon. This daemon collects from the main server the installation tasks and does the installation of resources, then notifies the main server of completion whenever a resource is configured, installed and available.

The data structure on the main server is the following:

A - Action: an action which can happen to provide a resource or account its usage
CP - Computer Partition: provides a URL to Access a Cloud Resource
RI - Resource Item: describes a resource
CI - Contract Item: describes the contract to attach the DL to (This is unclear still)
R - Resource: describes a type of cloud resource (ex. MySQL Table) is published on slapgrid.org
DL - Delivery Line: Describes an action happening on a resource item on a computer partition
D - Delivery: groups multiple Delivery Lines