* refactor: use repository in catalog
The catalog was refactored to use identifiers on everything, and store
everything in a consistent structure. This structure makes use of the
`Repository` type that holds a `SerdeVecMap` of Id to Resource, along
with the next Id, and a bi-map of Id to resource name.
The `Repository` type is used at each level of the catalog where a
resource is stored.
This simplified repeated logic for snapshot'ing, insert and update of
resources in the catalog, as well as accessor methods for getting by id
or name, and mapping names to ids and vice-versa.
In addition, the process for catalog batch verification and permit was
altered so that the permit process induces a retry if the catalog was
updated while the catalog batch function was producing the batch, i.e, if
the catalog sequence incremented while the caller was waiting for a permit.
This eliminated the need for verifying the catalog batch after it had been
generated, and allows for a single path to apply a catalog batch after it
has been persisted to object store.
This assumes that the generation of the catalog batch implies validity.
Irelevant tests were removed.
Last and Distinct cache's now rely more heavily on Ids, though the proc-
essing engine still needs to switch over to use Ids for starting/stopping
triggers.
This commit restores the old behavior we had where new tags can be added
to a schema. To do this we made tags nullable and brings us in line with
our other products. These changes were made in this PR:
https://github.com/influxdata/influxdb3_core/pull/41.
Changes to accomplish this new behavior were:
- Queries now do not return an empty string for null tags instead they
are returned as null, or in many formats not at all.
- References to v1 for parsing and validating lines were removed as we
only have one path for doing so these days shared amongst all the
write_lp endpoints.
- We fixed failing tests that expected us to not be able to have new
tags or depended on that functionality indirectly
- Tests had their snapshot files updated to reflect that tags are
nullable by default
- Behavior for making a schema and checking whether a column can be null
were updated in a separate repo and integrated here
- The series_key is updated whenever we get a new tag added to the
schema
- New tests were added to show that you can add a new tag and that the
series key is updated as part of that
With the above changes we can now allow tags to be added again by users
like they would expect, especially with v1 and v2 apis and Telegraf
plugins.
* chore: clean up runtime code for python setup
ccd5d22aab introduced working but temporary code for setting up the
python runtime environment. This cleans that up:
* refactor various find_python() functionality into virtualenv.rs
* refactor PYTHONHOME calculation to virtualenv.rs:find_python_install()
* adjust init_pyo3() to temporarily set PYTHONHOME based on
virtualenv.rs:find_python_install() as this is the only place it is
needed (indeed, venv activation scripts try to remove it)
Importantly, virtualenv.rs:find_python_install() tries to find the
python build standalone runtime based on a few heuristics. This function
could be improved in the fullness of time to, eg, be configured via a
build parameter.
Also, virtualenv.rs:find_python() can be used pre and post venv
activation. Before entering the venv, it will use find_python_install()
which is useful for things like setting up the initial venv. After
entering the venv, it will honor VIRTUAL_ENV (as set by
virtualenv.rs:initialize_venv()) to find python, which is important for
install packages with pip and having them installed into the venv.
* chore: update README_processing_engine.md for default venv
* chore: add bug reference for venv migrations with python minor releases
* chore: add security URLs to README_processing_engine.md
* chore: find_python_home() returns Option<PathBuf>. Thanks Jackson Newhouse
* fix: manually set sys.prefix, exec_prefix and sys.path
When activating a venv in the shell, sys.base_prefix and
sys.base_exec_prefix should be set to the installation location while
sys.prefix and sys.exec_prefix should be set to the venv dir.
Unfortunately, when initialize_venv() and init_pyo3() are called, we
can't use Py_InitializeFromConfig() to set any of these and certain
platforms are unable to find python-build-standalone. For now, we'll
temporarily set PYTHONHOME in init_pyo3() to the installation location
to make python work. By setting it at this point in the code, sys.prefix
and sys.exec_prefix end up also being set to the installation location,
which is fine when not under a venv, but is different from when entering
an venv.
To address this, in PYTHON_INIT.call_once() and when VIRTUAL_ENV is set
(which initialize_venv() will have set at this point), manually set
sys.prefix and sys.exec_prefix to what is in VIRTUAL_ENV.
Similarly, when activating a venv in the shell, sys.path is appended to
have the venv's site-packages dir. Previously we were setting PYTHONPATH
in initialize_venv() which ensures that the venv's site-packages dir is
in sys.path, but this ends up having the venv's site-packages dir first
in sys.path. To correct this, don't set PYTHONPATH any more and instead
adjust PYTHON_INIT.call_once() to append the venv's site-packages dir to
sys.path when VIRTUAL_ENV is set.
Finally, when exiting init_pyo3(), unconditionally unset PYTHONHOME when
VIRTUAL_ENV is set (like activation scripts do) and restore/unset when
it isn't.
Prior to these changes, all target incorrectly had the venv's
site-packages first in sys.path and OSX and Windows additionally had an
incorrect sys.prefix and sys.exec_prefix. With these initialization
changes in place, the runtime environment for the plugins is much closer
to that of a shell activated venv.
* feat(ci): fetch and configure for python-build-standalone binaries
* fix: make the process engine usable on windows
* feat(ci): build with python-build-standalone (and drop musl)
* fix(ci): set rpath on Linux and libpath on OSX in ci
* fix: set PYTHONHOME everywhere and PYTHONPATH on Windows
* chore(ci): update to use more recent ci-packager-next
* fix(ci): adjust validate to allow certain dynamically linked libraries
* chore: remove install_influxdb.sh (using install_influxdb3.sh instead)
* chore(install_influxdb3.sh): update for processing engine and release builds
* fix: temporarily use rpm --nodeps until compile with old GLIBC
* feat(ci): build docker with python-build-standalone
* chore: add README_processing_engine.md
* chore: add a few more details to README_processing_engine.md
* fix(ci): use patchelf --set-rpath
Not all patchelf versions support --add-rpath for appending to the
RPATH, but --set-path can be used with a colon-separated list. Use
--set-rpath first for maximum compatibility.
* chore: update README_processing_engine.md for standalone local builds
* fix(Dockerfile): also use patchelf --set-rpath
* chore: update code comment for accuracy
* chore: typos, grammar and formatting change in README_processing_engine.md
* chore: update README_processing_engine.md for Docker arm64 (thanks Jackson)
Partially fixes https://github.com/influxdata/influxdb/issues/24672
* move most HTTP req/resp types into `influxdb3_types` crate
* removes the use of locally-scoped request type structs from the `influxdb3_client` crate
* fix plugin dependency/package install bug
* it looks like the `DELETE` http method was being used where `POST` was expected for `/api/v3/configure/plugin_environment/install_packages` and `/api/v3/configure/plugin_environment/install_requirements`
This updates trigger creation to load the plugin file before creating the trigger.
Another small change is to make Github references use filenames and paths identical to what they would be in the plugin-dir. This makes it a little easier to have the plugins repo local and develop against it and then be able to reference the same file later with gh: once it's up on the repo.
This refactors plugins and triggers so that plugins no longer need to be "created". Since plugins exist in either the configured local directory or on the Github repo, a user now only needs to create a trigger and reference the plugin filename.
Closes#25876
* feat: first stab at locally updating parquet cache
closes: https://github.com/influxdata/influxdb/issues/25887
* refactor: use enums to separate out the modes
This commit introduced the `Immediate` and `Eventual` modes for
fulfilling the cache request. In immediate mode since the data is
readily available to be cached, we can avoid extra requests to object
store.
part of: https://github.com/influxdata/influxdb/issues/25887
This commit does a few key things:
- Removes the 72 hour query and write restrictions in Core
- Limits the queries to a default number of parquet files. We chose 432
as this is about 72 hours using default settings for the gen1
timeblock
- The file limit can be increased, but the help text and error message
when exceeded note that query performance will likely be degraded as
a result.
- We warn users to use smaller time ranges if possible if they hit this
query error
With this we eliminate the hard restriction we have in place, but
instead create a soft one that users can choose to take the performance
hit with. If they can't take that hit then it's recomended that they
upgrade to Enterprise which has the compactor built in to make
performant historical queries.
This updates plugins so that they will reload the code if the local file is modified. Github pugins continue to be loaded only once when they are initially created or loaded on startup.
This will make iterating on plugin development locally much easier.
Closes#25863
* feat: Add request plugin capability
Adds the request plugin type. Triggers can be bound to an API endpoint at /api/v3/engine/<path>. Requests will get yielded to the plugin with the query parameters, request parameters, and request body.
I didn't implement the test endpoint for this plugin type as it seems much more natural for users to save the file and make a new request. Once #25863 is done it'll make it very easy.
Closes#25862
* chore: fix spelling in error message
* feat(processing_engine): Add cron plugins and triggers to the processing engine.
* feat(processing_engine): switch from 'cron plugin' to 'schedule plugin', use TimeProvider.
* feat(processing_engine): add test for test scheduled plugin.
* feat: return better plugin execution errors
This sets up the framework for fleshing out more useful plugin execution errors that get returned to the user during testing. We'll also want to capture these for logging in system tables.
Also fixes a test that was broken in previous commit on time limits. Didn't show up because of the feature flag.
* fix: compile errors without system-py feature
This updates the create plugin API and CLI so that it doesn't take the pugin code, but instead takes a file name of a file that must be in the plugin-dir of the server. It returns an error if the plugin-dir is not configured or if the file isn't there.
Also updates the WAL and catalog so that it doesn't store the plugin code directly. The code is read from disk one time when the plugin runs.
Closes#25797
* feat: introduce num wal files to keep
This commit allows a configurable number of wal files to be left behind
in OS. This is necessary as enterprise replicas rely on these files.
closes: https://github.com/influxdata/influxdb/issues/25788
* refactor: address PR feedback
* refactor: address PR comment
This allows the user to specify arguments that will be passed to each execution of a wal plugin trigger. The CLI test was updated to check this end to end.
Closes#25655
This updates the WAL so that it can have new file notifiers added that will get updated when the wal flushes. The processing engine now implements the WALNotifier trait.
I've updated the CLI test for creating a trigger to run and end-to-end test that defines a plugin, creates a trigger, writes data into the database, triggering the plugin, which writes summary statistics back into the database in a different table. The test queries the destination table to confirm that the plugin worked.
* feat: Update WAL plugin for new structure
This ended up being a very large change set. In order to get around circular dependencies, the processing engine had to be moved into its own crate, which I think is ultimately much cleaner.
Unfortunately, this required changing a ton of things. There's more testing and things to add on to this, but I think it's important to get this through and build on it.
Importantly, the processing engine no longer resides inside the write buffer. Instead, it is attached to the HTTP server. It is now able to take a query executor, write buffer, and WAL so that the full range of functionality of the server can be exposed to the plugin API.
There are a bunch of system-py feature flags littered everywhere, which I'm hoping we can remove soon.
* refactor: PR feedback
Trevor Hilton
Jackson Newhouse
Michael Gattozzi
Jamie Strandboge
wayne
Paul Dix
praveen-influx