This recipe creates and configures a Zope instance in parts. (Despite its name it nowadays only works for Zope 4+.) It also installs a control script, which is like zopectl, in the bin/ directory. The name of the control script is the name of the part in buildout. By default various runtime and log information will be stored inside the var/ directory.
You can use it with a part like this:
[instance] recipe = plone.recipe.zope2instance user = admin:admin http-address = 8080 eggs = my.distribution zcml = my.distribution
Contents
- Introduction
- Options
- The generated control script
- Reporting bugs or asking questions
- eggs
- The list of distributions you want to make available to the instance.
- zcml
- Install ZCML slugs for the distributions listed, separated by whitespace. You
can specify the type of slug by appending '-' and the type of slug you want
to create. Some examples:
my.distributionmy.distribution-meta. - http-address
- Set the address of the HTTP server. Can be either a port or a socket address. Defaults to 0.0.0.0:8080.
- ip-address
- The default IP address on which Zope's various server protocol implementations will listen for requests. If this is unset, Zope will listen on all IP addresses supported by the machine. This directive can be overridden on a per-server basis in the servers section. Defaults to not setting an ip-address. Used for ZServer only, not WSGI.
- threads
- Specify the number of worker threads used to service requests. The default is 4 for WSGI (since this is the waitress default) and 2 for ZServer.
- zodb-cache-size
- Set the ZODB cache size, i.e. the number of objects which the ZODB cache will try to hold. Defaults to 30000.
- zserver-threads
- Deprecated, use threads instead. Specify the number of threads that Zope's ZServer web server will use to service requests. The recipes default is 2. Used for ZServer only, not WSGI.
- environment-vars
Define arbitrary key-value pairs for use as environment variables during Zope's run cycle. Example:
environment-vars = TZ US/Eastern zope_i18n_allowed_languages en zope_i18n_compile_mo_files true
- initialization
- Specify some Python initialization code to include within the generated
sitecustomize.pyscript (Buildout >= 1.5) or within the instance script (Buildout < 1.5). This is very limited. In particular, be aware that leading whitespace is stripped from the code given. added in version 4.2.14 - wsgi
By default this recipe creates a Python script that uses
waitressas a WSGI server. When running Python 2 you can disable WSGI and use ZServer by settingwsgi = offand including ZServer in theeggsspecification list. Example:wsgi = off eggs = ... ZServer
You can use other PasteDeploy-compatible WSGI servers by passing a path to a WSGI configuration file here and including the WSGI server's egg in the
eggsspecification. Example:wsgi = ${buildout:directory}/etc/gunicorn.ini eggs = ... gunicornThe WSGI configuration file will not be created for you in this case, unlike the built-in
waitresssupport. You have to provide it yourself.- max-request-body-size
- Specify the maximum request body size in bytes The default is 1073741824 (since this is the waitress default)
Please refer to https://pypi.org/project/plone.resource for more details and setup instructions.
- resources
Specify a central resource directory. Example:
resources = ${buildout:directory}/resources
- locales
Specify a locales directory. Example:
locales = ${buildout:directory}/locales
This registers a locales directory with extra or different translations.
If you want to override a few translations from the plone domain in the
English language, you can add a en/LC_MESSAGES/plone.po file in this
directory, with standard headers at the top, followed by something like
this:
#. Default: "You are here:" msgid "you_are_here" msgstr "You are very welcome here:"
Translations for other message ids are not affected and will continue to work.
- verbose-security
- Set to on to turn on verbose security (and switch to the Python security implementation). Defaults to off (and the C security implementation).
- debug-exceptions
- WSGI only: set to
onto disable exception views includingstandard_error_message. Exceptions other thanUnauthorizedorConflictErrorcan then travel up into the WSGI stack. Use this option if you want more convenient error debugging offered by WSGI middleware such as the werkzeug debugger. See the Zope WSGI documentation for examples. - profile
- Set to
onenables repoze.profile. Defaults tooff, If switched on there are further options prefixed withprofile_to configure it as below. You will need to add the repoze.profile package, either by adding it to your eggs section directly or by using the extra plone.recipe.zope2instance[profile]. - profile_log_filename
- Filename of the raw profile data.
Default to
profile-SECTIONNAME.raw. This file contains the raw profile data for further analysis. - profile_cachegrind_filename
- If the package
pyprof2calltreeis installed, another file is written. It is meant for consumation with any cachegrind compatible application. Defaults tocachegrind.out.SECTIONNAME. - profile_discard_first_request
- Defaults to
true. See repoze.profile docs for details. - profile_path
- Defaults to
/__profile__. The path for through the web access to the last profiled request. - profile_flush_at_shutdown
- Defaults to
true. See repoze.profile docs for details. - profile_unwind
- Defaults to
false. See repoze.profile docs for details.
If you have only one application process, it can open the database files directly without running a database server process.
- file-storage
- The filename where the ZODB data file will be stored. Defaults to ${buildout:directory}/var/filestorage/Data.fs.
- blob-storage
- The name of the directory where the ZODB blob data will be stored, defaults to ${buildout:directory}/var/blobstorage.
If you want multiple application processes you need to run a separate database server process and connect to it, either via ZEO or RelStorage.
- zeo-address
- Set the address of the ZEO server. Defaults to 8100. You can set more than one address (white space delimited). Alternative addresses will be used if the primary address is down.
- zeo-client
- Set to 'on' to make this instance a ZEO client. In this case, setting the zeo-address option is required, and the file-storage option has no effect. To set up a ZEO server, you can use the plone.recipe.zeoserver recipe. Defaults to 'off'.
- blob-storage
- The location of the blob zeocache, defaults to var/blobcache. If shared-blob is on it defaults to ${buildout:directory}/var/blobstorage.
- shared-blob
- Defaults to off. Set this to on if the ZEO server and the instance have access to the same directory. Either by being on the same physical machine or by virtue of a network file system like NFS. Make sure this instances blob-storage is set to the same directory used for the ZEO servers blob-storage. In this case the instance will not stream the blob file through the ZEO connection, but just send the information of the file location to the ZEO server, resulting in faster execution and less memory overhead.
- zeo-client-read-only-fallback
- A flag indicating whether a read-only remote storage should be acceptable as a fallback when no writable storages are available. Defaults to false.
- read-only
- Set zeo client as read only added in version 4.2.12
You need to activate ZEO auth on the server side as well, for this to work. Without this anyone that can connect to the database servers socket can read and write arbitrary data.
- zeo-username
- Enable ZEO authentication and use the given username when accessing the ZEO server. It is obligatory to also specify a zeo-password.
- zeo-password
- Password to use when connecting to a ZEO server with authentication enabled.
- zeo-realm
- Authentication realm to use when authentication with a ZEO server. Defaults to 'ZEO'.
Please refer to https://pypi.org/project/RelStorage for more details and setup instructions.
- rel-storage
Allows to set a RelStorage instead of a FileStorage.
Contains settings separated by newlines, with these values:
- type: any database type supported (postgresql, oracle, mysql)
- RelStorage specific keys, like cache-servers and poll-interval
- all other keys are passed on to the database-specific RelStorage adapter.
Example:
rel-storage = type oracle dsn (DESCRIPTION=(ADDRESS=(HOST=s01))(CONNECT_DATA=(SERVICE_NAME=d01))) user tarek password secret
In most cases you don't need to adjust any of this, you might want to adjust log levels or configure mailinglogger.
- event-log
- The filename of the event log. Defaults to ${buildout:directory}/var/log/${partname}.log Setting this value to 'disable' will make the <eventlog> section to be omitted, disabling logging events by default to a .log file.
- event-log-level
- Set the level of the console output for the event log. Level may be any of CRITICAL, ERROR, WARN, INFO, DEBUG, or ALL. Defaults to INFO.
- event-log-max-size
- Maximum size of event log file. Enables log rotation. Used for ZServer only, not WSGI.
- event-log-old-files
- Number of previous log files to retain when log rotation is enabled. Defaults to 1. Used for ZServer only, not WSGI.
- event-log-custom
- A custom section for the eventlog, to be able to use another event logger than logfile. Used for ZServer only, not WSGI.
- mailinglogger
A mailinglogger section added into the event log. Used for ZServer only, not WSGI. Example snippet:
<mailing-logger> level error flood-level 10 smtp-server smtp.mydomain.com from [email protected] to [email protected] subject [My domain error] [%(hostname)s] %(line)s </mailing-logger>
You will need to add mailinglogger to your buildout's egg section to make this work.
- access-log, z2-log
- The filename for the Z2 access log. Defaults to var/log/${partname}-Z2.log (var/log/${partname}-access.log) for WSGI). You can disable access logging by setting this value to 'disable'. For ZServer this will omit the <logger access> section in zope.conf. For WSGI, the logging handler will be a NullHandler.
- access-log-level, z2-log-level
- Set the log level for the access log. Level may be any of CRITICAL, ERROR, WARN, INFO, DEBUG, or ALL. Defaults to WARN (INFO for WSGI).
- access-log-max-size
- Maximum size of access log file. Enables log rotation. Used for ZServer only, not WSGI.
- access-log-old-files
- Number of previous log files to retain when log rotation is enabled. Defaults to 1. Used for ZServer only, not WSGI.
- access-log-custom
- Like event-log-custom, a custom section for the access logger, to be able to use another event logger than logfile. Used for ZServer only, not WSGI.
- sentry_dsn
- Provide a Sentry DSN here to enable basic Sentry logging documented in https://docs.sentry.io/platforms/python/logging/. You will need to add the Python Sentry SDK, either by adding it to your eggs section directly or by adding plone.recipe.zope2instance[sentry]. Available for WSGI only.
- sentry_level
- Set the logging level for Sentry breadcrumbs. Available for WSGI only.
- sentry_event_level
- Set the logging level for Sentry events. Available for WSGI only.
- sentry_ignore
- Set the (space separated list of) logger names that are ignored by Sentry. Available for WSGI only.
- sentry_max_value_length
- Set the maximum size of traceback messages sent to Sentry. If your tracebacks get truncated, increase this above the sentry-sdk default of 1024. Available for WSGI only.
For more complex logging configuration, the zope2instance recipe exposes the underlaying logging.handlers functionality through the access-log-handler and event-log-handler configuration options. This allows you to configure an arbitrary logging handler for Python as defined in here.
The supplementary options event-log-args, event-log-kwargs and access-log-args, access-log-kwargs can be used for passing positional and keyword arguments to the constructor of the underlaying handler.
- access-log-handler
The (dotted) name of an importable Python logging handler like logging.handlers.RotatingFileHandler.
Default: FileHandler
- access-log-args
A python tuple which usually refers to the logging filename and opening mode of the file like ("access.log", "a"). Note that you a Python tuple with only one element (e.g. only the filename) must have a trailing comma like ("access.log", ) The access-log-args is used to specify the positional parameters for the logging handler configured through access-log-handler.
Default: (r"access.log", "a")
- access-log-kwargs
A python dictionary used for passing keyword argument for the logging handler configured through access-log-handler e.g. {"when": "h", "interval": 1}.
Default: {}
- event-log-handler
- Same as access-log-handler but for the configuration of the event log of Plone.
- event-log-args
- Same as access-log-args but for the configuration of the event log of Plone.
- event-log-kwargs
- Same as access-log-kwargs but for the configuration of the event log of Plone.
- wsgi-logging-ini-template
By default plone.recipe.zope2instance uses a hard-coded logging template for the generated WSGI configuration in parts/<partname>/etc/wsgi.ini. The template is defined as wsgi_logging_ini_template variable within the recipe.py file.
You can override the template with a custom wsgi_logging.ini file using this option. All other default wsgi options will be untouched.
Example:
wsgi-logging-ini-template = /path/to/wsgi_logging.ini
This option cannot be used together with the wsgi-ini-template option (see Advanced options: wsgi-ini-template)
This example uses a RotatingFileHandler https://docs.python.org/3/library/logging.handlers.html#rotatingfilehandler which rotates the access log when it becomes larger than 10 MB while keeping seven copies:
access-log-handler = logging.handlers.RotatingFileHandler
access-log-args = (r"access.log", "a")
access-log-kwargs = {"maxBytes": 10000000, "backupCount": 7}
This example uses a TimedRotatingFileHandler https://docs.python.org/3/library/logging.handlers.html#timedrotatingfilehandler for rotating the event log every 24 hours or one day:
event-log-handler = logging.handlers.TimedRotatingFileHandler
event-log-args = (r"event.log", )
event-log-kwargs = {"when": "D", "interval": 1}
- products
- A list of paths where Zope 2 products are installed. The first path takes precedence in case the same product is found in more than one directory. Zope 2 products are deprecated and won't work any longer in a future version of Zope/Plone.
- extra-paths
- A list of paths where additional Python packages are installed. The paths are searched in the given order after all egg and products paths.
- site-zcml
- If you want a custom site.zcml file, put its content here. If this option is used the zcml and zcml-additional options are ignored.
- zcml-additional
- Extra ZCML statements that should be included in the generated site.zcml file.
- zeo-client-cache-size
- Set the size of the ZEO client cache. Defaults to '128MB'. The ZEO cache is a disk based cache shared between application threads. It is stored either in temporary files or, in case you activate persistent cache files with the option client (see below), in the folder designated by the zeo-var option.
- zeo-client-client
- Set the persistent cache name that is used to construct the cache filenames. This enables the ZEO cache to persist across application restarts. Persistent cache files are disabled by default.
- zeo-client-blob-cache-size
- Set the maximum size of the ZEO blob cache, in bytes. If not set, then the cache size isn't checked and the blob directory will grow without bound.
- zeo-client-blob-cache-size-check
- Set the ZEO check size as percent of zeo-client-blob-cache-size (for example, 10 for 10%). The ZEO cache size will be checked when this many bytes have been loaded into the cache. Defaults to 10% of the blob cache size. This option is ignored if shared-blob is enabled.
- zeo-client-drop-cache-rather-verify
- Indicates that the cache should be dropped rather than verified when the verification optimization is not available (e.g. when the ZEO server restarted). Defaults to 'False'.
- zeo-storage
- Set the storage number of the ZEO storage. Defaults to '1'.
- zeo-var
- Used in the ZEO storage snippets to configure the ZEO var folder, which is used to store persistent ZEO client cache files. Defaults to the system temporary folder.
- wsgi-ini-template
By default plone.recipe.zope2instance uses a hard-coded template for the generated WSGI configuration in parts/<partname>/etc/wsgi.ini. The template is defined as wsgi_ini_template variable within the recipe.py file.
You can override the template with a custom template file using this option.
Example:
wsgi-ini-template = /path/to/wsgi_template.ini
The available variables for variable substition can be found within the existing template (see above).
- asyncore-use-poll
- By default false. If you want the waitress.asyncore.loop flag to use poll() instead of the default select() set to true.
- before-storage
Wraps the base storage in a "before storage" which sets it in read-only mode from the time given (or "now" for the current time).
This option is normally used together with demo-storage for a normally running site in order for changes to be made to the database.
- client-home
- Sets the clienthome for the generated instance. Defaults to ${buildout:directory}/var/<name of the section>.
- clear-untrusted-proxy-headers
- This tells Waitress to remove any untrusted proxy headers
("Forwarded", "X-Forwarded-For", "X-Forwarded-By",
"X-Forwarded-Host", "X-Forwarded-Port", "X-Forwarded-Proto").
The default in waitress 1 is false, but waitress 2 changes this to true.
We explicitly default to false.
When you set it to true, you may need to set other
wsgi.inioptions liketrusted_proxy_headersandtrusted_proxy. Setting those is not supported by the recipe yet. Used for WSGI only, not ZServer. - default-zpublisher-encoding
- This controls what character set is used to encode unicode data that reaches ZPublisher without any other specified encoding. This defaults to 'utf-8'. Plone requires this to be set to utf-8.
- demo-storage
If 'on' it enables the demo storage. By default, this is a memory-based storage option; changes are not persisted (see the demo-file-storage option to use a persistent storage for changes made during the demonstration).
To use with a base storage option configured with a blob-storage, you must set a demo-blob-storage.
- demo-file-storage
- If provided, the filename where the ZODB data file for changes committed during a demonstration will be stored.
- demo-blob-storage
If provided, the name of the directory where demonstration ZODB blob data will be stored.
This storage may be connected to a demonstration file storage, or used with the default memory-based demo storage (in this case you might want to use a temporary directory).
- storage-wrapper
- Template for arbitrary configuration to be wrapped around the main storage. %s will be replaced with the existing storage configuration.
- effective-user
- The name of the effective user for the Zope process. Defaults to not setting an effective user.
- enable-product-installation
- Enable the persistent product registry by setting this to
on. By default the registry is turnedoff. Enabling the registry is deprecated. - ftp-address
- Give a port for the FTP server. This enables the FTP server. Used for ZServer only, not WSGI.
- http-force-connection-close
- Set to on to enforce Zope to set
Connection: close header. This is useful if for example a 304 leaves the connection open with Varnish in front and Varnish tries to reuse the connection. - http-fast-listen
- Set to off to defer opening of the HTTP socket until the end of the Zope startup phase. Defaults to on.
- icp-address
- Give a port for the ICP server. This enables the ICP server. Used for ZServer only, not WSGI.
- import-directory
- Used to configure the import directory for instance. Defaults to <client-home>/import.
- port-base
- Offset applied to the port numbers used for ZServer configurations. For example, if the http-server port is 8080 and the port-base is 1000, the HTTP server will listen on port 9080. This makes it easy to change the complete set of ports used by a Zope server process. Zope defaults to 0.
- python-check-interval
- An integer telling the Python interpreter to check for asynchronous events every number of instructions. This affects how often thread switches occur. Defaults to 1000.
- relative-paths
- Set this to true to make the generated scripts use relative paths. You can also enable this in the [buildout] section.
- scripts
Add this parameter with no arguments to suppress script generation. Otherwise (i.e. without this parameter), scripts for packages added to the eggs parameter will be generated. You may also configure per package. E.g.:
[instance] recipe = plone.recipe.zope2instance eggs = Plone mr.migrator my.package scripts = my_package_script
In the above example, only my_package_script will be generated. Keep in mind that the egg containing the script (
my.packagein the example) must be listed explicitly in the eggs option, even if it is a dependency of an already listed egg.- template-cache
- Used to configure the cache for page-template files. Chameleon will write compile page-templates into this directory and use it as a cache. See https://chameleon.readthedocs.io/en/latest/configuration.html for more info. Valid options are off or on or a directory-location. Defaults to ${buildout:directory}/var/cache (it also confirms to what var is set to).
- var
- Used to configure the base directory for all things going into var. Defaults to ${buildout:directory}/var.
- webdav-address
- Give a port for the WebDAV server. This enables the WebDAV server. Used for ZServer only, not WSGI.
- webdav-force-connection-close
- Valid options are off and on. Defaults to off. Used for ZServer only, not WSGI.
- pipeline
The main application pipeline served by the wsgi server. By default the pipeline is:
translogger egg:Zope#httpexceptions zope
The
transloggerline in the pipeline will be removed ifz2-logis set todisabledor if it is not set andaccess-logis set todisabled(case insensitive). Used for WSGI only, not ZServer.- zlib-storage
Adds support for file compression on a file storage database. The option accepts the values 'active' (compress new records) or 'passive' (do not compress new records). Both options support already compressed records.
You can use the 'passive' setting while you prepare a number of connected clients for compressed records.
- zodb-cache-size-bytes
- Set the ZODB cache sizes in bytes. This feature is still experimental.
- zodb-temporary-storage
- If given Zope's default temporary storage definition will be replaced by the lines of this parameter. If set to "off" or "false", no temporary storage definition will be created. This prevents startup issues for basic Zope 4 sites as it does not ship with the required packages by default anymore.
- zope-conf
- A relative or absolute path to a zope.conf file. If this is given, many of the options in the recipe will be ignored.
- zope-conf-imports
You can define custom sections within zope.conf using the ZConfig API. But, in order for Zope to understand your custom sections, you'll have to import the python packages that define these custom sections using %import syntax.
Example:
zope-conf-imports = mailinglogger eea.graylogger
- zope-conf-additional
Give additional lines to zope.conf. Make sure you indent any lines after the one with the parameter.
Example:
zope-conf-additional = locale fr_FR http-realm Slipknot
- zopectl-umask
Manually set the umask for the zopectl process.
Example:
zopectl-umask = 002
- http-header-max-length
Manually set the maximum size of received HTTP header being processed by Zope. The request is discarded and considered as a DoS attack if the header size exceeds this limit. Default: 8192. Used for ZServer only, not WSGI.
Example:
http-header-max-length = 16384
On the windows platform the bin/instance script as described below will not be generated, because it uses a Unix specific implementation.
To run Plone start it with:
.\bin\runwsgi.exe -v .\parts\etc\wsgi.ini
Or for development in debug mode use:
.\bin\runwsgi.exe -v .\parts\etc\wsgi.ini
The documentation for the extended Zope control script below does not apply.
The extended Zope control script installed by this recipe, usually bin/instance by convention, offers a debug command and another run command. The debug command starts an interactive Python prompt with the Zope application available via the app name. Similarly, the run command accepts a Python script as an argument that will be run under the same conditions.
These commands have also been extended to set up a more complete environment. Specifically, these commands set up a REQUEST, log in the AccessControl.SpecialUsers.system user, and may traverse to an object, such as a CMF portal. This environment set up is controlled with following options:
-R/--no-request -- do not set up a REQUEST.
-L/--no-login -- do not login the system user.
-O/--object-path <path> -- Traverse to <path> from the app
and make available as `obj`.
Note that these options must come before the script name, e.g. bin/instance -RLOPlone/front-page debug
The console command is similar to the fg command, but it does not create a subprocess to start up Zope. This is useful for two use cases. First, the supervisor program, to supervise long running processes like a Zope, require the process not to fork away, so that supervisor can control it. Second, IDEs like WingIDE and PyCharm support debugging running processes from within. For this to work, the process should also not fork away.
Third-party distributions may add additional commands to the control script by installing a 'plone.recipe.zope2instance.ctl' entry point. For example, an egg called MyDist could include a module called mymodule with the following custom command:
def foo(self, *args)
"""Help message here"""
print 'foo'
It would then install the foo method as a command for the control script using the following entry point configuration in setup.py:
entry_points=""" [plone.recipe.zope2instance.ctl] foo = mymodule:foo """
This would allow invoking the foo method by running bin/instance foo (assuming the instance control script was installed by a buildout part called instance.) The entry point is invoked with the following parameters:
- self
- An instance of plone.recipe.zope2instance.ctl.AdjustedZopeCmd.
- args
- Any additional arguments that were passed on the command line.
- the
restartcommand will not function reliably if you run the buildout while the Zope instance is still running. In those cases, always usestopfollowed bystartto restart the Zope instance.
Please use the bug tracker in this repository at https://github.com/plone/plone.recipe.zope2instance/issues for questions and bug reports.