mirror of
https://github.com/valitydev/salt.git
synced 2024-11-08 01:18:58 +00:00
336fbb03bc
* Fix master hanging after a request from minion with removed key. (#33333) * ZMQ monitor for MWorker connections. * Reauth minion if the key was removed on the master side. * Allow concurrency mode in state runs if using sudo (#33325) Closes #30130 * Disambiguate non-exact matches when checking if sysv service is enabled (#33324) Fixes #33323 * remove redundant, incorrect sudo_runas config documentation (#33318) * remove sudo_runas documentation `sudo_runas` was renamed to `sudo_user` and the documentation was not updated accordingly. * conf/minion: update sudo_user description The description from sudo_runas was better. * import ps from psutil_compat in beacons (#33334) * beacons.network_info: import gate psutil * beacons.ps: import gate psutil * Add docs for mine_functions config var (#33326) * Add docs for mine_functions config var * Note that mine_enabled essentially just doesn't add the mine update function to the scheduler. * Bp 28467 calm mine (#33327) * make minion mine update behavior more configurable * Add docs for mine_functions config var * Remove config dup from mine config options Refs #28467 * 2015.8 does not have _DFLT_MULTIPROCESSING_MODE * This won't be in until 2015.8.10. * Fix network.managed for windows (#33312) * Fix some link errors in the test writing tutorial (#33347) * Describes parameters in register_instances function (#33339) * Fix UnboundLocalError in git.latest (#33340) Resolves #32260. * Expanded documentation for boto_elb state and module (#33341) * Describes what happens when the CNAME parameter is given. * Describes what the recognized attributes are for for ELBs. * Properly detect newer Linux Mint distros (#33359) * Properly detect newer Linux Mint distros LMDE 2 and Linux Mint 17.3 changed the DISTRIB_ID in /etc/lsb-release to ``LinuxMint``, breaking OS detection for these distros. This commit fixes that by adding an entry to the OS_NAME_MAP in the core grains. * Remove LinuxMint os_family from aptpkg.py It is no longer necessary as the distro is now detected properly, which will lead to an os_family of Debian. * Update job_cache and keep_jobs docs to be more specific to their behavior (#33328) * Update job_cache and keep_jobs docs to be more specific to their behavior Also fixed a bug discovered when investigating job_cache/keep_jobs functionality where the jid directory and files were removed by the cache cleaner, but not the original jid clash detection directory created in /var/cache/salt/master/jobs/. Fixes #29286 * Add testcase for the changes in the local_cache.clean_old_jobs func * Mark tests as destructive * Put destructive test decorator in correct location * Remove mentions of windows not supporting pkgs param (#33361) Fixes #33313 * Updates docs version to 2015.8.9 Adds note regarding the os grain on Mint Linux Adds an FAQ regarding grains that change due to upstream changes * revved 2015.8 branch to .9 in version selector * Add initscripts, SystemD service units and environment files for Debian (#32857) * Add note to docs about api settings for Hipchat API v2 (#33365) Fixes #27779 * Add win_pkg to list of modules that support "version" in pkg.installed (#33362) Fixes #32913 * Add note about name parameter in git_pillar docs (#33369) Fixes #27737 * Better YAML syntax error handling (#33375) Closes #26574 * Improve doc clarity for disable_modules documentation (#33379) * Improve doc clarity for disable_modules documentation * Additional clarification on blacklisted name * maintain the fallabck because I am totally sick of this crap * blast, put the try/except int he right place * restore whitespace * Fix traceback in logging for config validation (#33386) * 2015.8.10 release notes * Sync pillarstack to latest upstream version (#33391) * Don't lay down all available opts (#33385) * Don't lay down all available opts * We need at least one opt in there * Condense defaults * Put the default hash type back
73 lines
2.7 KiB
ReStructuredText
73 lines
2.7 KiB
ReStructuredText
.. _managing_the_job_cache:
|
|
|
|
======================
|
|
Managing the Job Cache
|
|
======================
|
|
|
|
The Salt Master maintains a job cache of all job executions which can be
|
|
queried via the jobs runner. This job cache is called the Default Job Cache.
|
|
|
|
.. _default_job_cache:
|
|
|
|
Default Job Cache
|
|
=================
|
|
|
|
A number of options are available when configuring the job cache. The default
|
|
caching system uses local storage on the Salt Master and can be found in the
|
|
job cache directory (on Linux systems this is typically
|
|
``/var/cache/salt/master/jobs``). The default caching system is suitable for most
|
|
deployments as it does not typically require any further configuration or
|
|
management.
|
|
|
|
The default job cache is a temporary cache and jobs will be stored for 24
|
|
hours. If the default cache needs to store jobs for a different period the
|
|
time can be easily adjusted by changing the `keep_jobs` parameter in the
|
|
Salt Master configuration file. The value passed in is measured via hours:
|
|
|
|
|
|
.. code-block:: yaml
|
|
|
|
keep_jobs: 24
|
|
|
|
Reducing the Size of the Default Job Cache
|
|
------------------------------------------
|
|
|
|
The Default Job Cache can sometimes be a burden on larger deployments (over 5000
|
|
minions). Disabling the job cache will make previously executed jobs unavailable
|
|
to the jobs system and is not generally recommended. Normally it is wise to make
|
|
sure the master has access to a faster IO system or a tmpfs is mounted to the
|
|
jobs dir.
|
|
|
|
However, you can disable the :conf_master:`job_cache` by setting it to ``False``
|
|
in the Salt Master configuration file. Setting this value to ``False`` means that
|
|
the Salt Master will no longer cache minion returns, but a JID directory and ``jid``
|
|
file for each job will still be created. This JID directory is necessary for
|
|
checking for and preventing JID collisions.
|
|
|
|
The default location for the job cache is in the ``/var/cache/salt/master/jobs/``
|
|
directory.
|
|
|
|
Setting the :conf_master:`job_cache`` to ``False`` in addition to setting
|
|
the :conf_master:`keep_jobs` option to a smaller value, such as ``1``, in the Salt
|
|
Master configuration file will reduce the size of the Default Job Cache, and thus
|
|
the burden on the Salt Master.
|
|
|
|
.. note::
|
|
|
|
Changing the ``keep_jobs`` option sets the number of hours to keep old job
|
|
information and defaults to ``24`` hours. Do not set this value to ``0`` when
|
|
trying to make the cache cleaner run more frequently, as this means the cache
|
|
cleaner will never run.
|
|
|
|
|
|
Additional Job Cache Options
|
|
============================
|
|
|
|
Many deployments may wish to use an external database to maintain a long term
|
|
register of executed jobs. Salt comes with two main mechanisms to do this, the
|
|
master job cache and the external job cache.
|
|
|
|
See :ref:`Storing Job Results in an External System <external-master-cache>`.
|
|
|
|
|