-
Notifications
You must be signed in to change notification settings - Fork 24.4k
Redis Function Libraries #10004
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Redis Function Libraries #10004
Conversation
Libraries purpose is to provide a better code sharing between functions by allowing
to create multiple functions in a single command. Functions that were created together
can safely share code between each other without worrying about compatibility issues
and versioning.
Creating a new library is done using 'FUNCTION LOAD' command (full API is described below)
This PR introduces a new struct called libraryInfo, libraryInfo holds information
about a library:
* name - name of the library
* engine - engine used to create the library
* code - library code
* description - library description
* functions - the functions exposed by the library
When Redis gets the `FUNCTION LOAD` command it creates a new empty libraryInfo.
Redis passes the `CODE` to the relevant engine alongside the empty libraryInfo.
As a result, the engine will create one or more functions by calling 'libraryCreateFunction'.
The new funcion will be added to the newly created libraryInfo. So far Everything is
happening locally on the libraryInfo so it is easy to abort the operation (in case
of an error) by simply freeing the libraryInfo. After the library info is fully
constructed we start the joining phase by which we will join the new library to
the other libraries currently exist on Redis. The joining phase make sure
there is no function collision and add the library to the librariesCtx (renamed
from functionCtx). LibrariesCtx is used all around the code in the exact same way
as functionCtx was used (with respect to RDB loading, replicatio, ...). The
only difference is that apart from function dictionary (maps function name to
functionInfo object), the librariesCtx contains also a libraries dictionary that
maps library name to libraryInfo object.
`FUNCTION LOAD <ENGINE> <LIBRARY NAME> [REPLACE] [DESCRIPTION <DESCRIPTION>] <CODE>`
Create a new library with the given parameters:
* ENGINE - REPLACE Engine name to use to create the library.
* LIBRARY NAME - The new library name.
* REPLACE - If the library already exists, replace it.
* DESCRIPTION - Library description.
* CODE - Library code.
Return "OK" on success, or error on the following cases:
* Library name already taken and REPLACE was not used
* Name collision with another existing library (even if replace was uses)
* Library registration failed by the engine (usually compilation error)
`FUNCTION LIST [LIBRARYNAME <LIBRARY NAME PATTERN>] [WITHCODE]`
Command was modified to also allow getting libraries code (so `FUNCTION INFO` command
is no longer needed and removed). In addition the command gets an option argument,
`LIBRARYNAME` allows you to only get libraries that match the given `LIBRARYNAME`
pattern. By default, it returns all libraries.
* FUNCTION CREATE
* FUNCTION INFO
When the Lua engine gets the code given on `FUNCTION LOAD` command, it immediately runs
it, we call this run the loading run. Loading run is not a usual script run, it is not
possible to invoke any Redis command from within the load run. The only Redis API provided
to the load run is:
* redis.log
* redis.sha1hex
* redis.error_reply
* redis.status_reply
* redis.register_function
The loading run purpose is to register functions using the new `redis.register_function` API.
Any attempt to use any other API will result in an error.
`redis.register_function(<function_name>, <callback>, [<description>])`
This new API allows users to register a new function that will be linked to the newly created library.
This API can only be called during the load run (see definition above). Any attempt to use it
outside of the load run will result in an error. The parameters pass to the API are:
* function_name - Function name (must be a Lua string)
* callback - Lua function object that will be called when the function is invokes using fcall/fcall_ro
* description - Function description, optional (must be a Lua string).
The following example creates a library called `lib` with 2 functions, `f1` and `f1`, returns 1 and 2 respectively:
```
local function f1(keys, args)
return 1
end
local function f2(keys, args)
return 2
end
redis.register_function('f1', f1)
redis.register_function('f2', f2)
```
Notice: Unlike `eval`, functions inside a library get the KEYS and ARGV as arguments to the functions
and not as globals.
|
@MeirShpilraien i've set this to resolve the issue that describes it, assuming that all the ideas specified there are handled here. |
|
@oranagra function flags was missing, added it to todo list. |
1. Added WITHCODE option that allows to get library code 2. Added LIBRARY name that allows to get only libraries that matches the given library name pattern
1. Added libraries memory overhead to total functions overhead. 2. Added number of libraries to 'info memory'
oranagra
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
One concern i have about naming, is the use of the term "library" without a context.
in the past i argued the same about "engine".
the term "function" is a lost cause (in the context of redis, a "function" is a "script function".
but maybe "library" should be called "functions library" (or "functionsLib")?
same as we called "engine" a "script engine"
Co-authored-by: Oran Agra <oran@redislabs.com>
1. Nullify iterators so in case we have other (later) goto done, they won't break.
2. On librariesCtx join, used tmp list to keep old libraries instead on tmp libraraiesCtx.
3. On libraryJoin, if replace is used, do not search for collids libraries because we
there is no such.
4. Avoid nullify dictionary entry variable.
5. Added a comment about functions name collision to FUNCTION HELP
6. Better naming (li -> old_li)
7. Documented 'luaEngineLoadHook'
8. Added argument description to function-list.json
9. Avoid using setDeferredArrayLen if possible
10. Change error message to be generic to scripts (when running Lua api outside of script invocation).
11. Used global if statement to decide which argument to use to invoke 'lua_pcall'
|
@oranagra fixed the comments.
I think its ok to call it just library, but I also thought its ok to just say engine so it make sense we disagree :)
Let me know if you are OK with it. |
|
@MeirShpilraien i'm not sure if just renaming these two is enough. maybe there are other functions (procedures) that should be renamed, comments that should be updated, variable names or command arguments, etc. i suppose that in the context of a FUNCTION command, the term "library" is ok. |
Co-authored-by: Oran Agra <oran@redislabs.com>
|
@oranagra let me know what you think on the last PR (renaming) |
|
LGTM (only looked at the commit comment)
|
|
@oranagra I actually look everywhere where any function API is used and then find out what need to be renamed. Will do another round to make sure I did not miss anything. |
|
@redis/core-team this is ready for merge. please read the top comment and approve |
The PR add Redis Programability documentation: * Documentation to all FUNCTION commands (notice that FUNCTION INFO is not documented because its about to be dropped): * FUNCTION LOAD (current named FUNCTION CREATE but will be renamed after [this](redis/redis#10004) will be merged) * FUNCTION DELETE * FUNCTION DUMP * FUNCTION FLUSH * FUNCTION KILL * FUNCTION LIST * FUNCTION RESTORE * FUNCTION FLUSH * Introduction to Redis Functions page * Rearrangement of `EVAL` command page * All the information on `EVAL` command page was moved and split between 2 pages: * Introduction to Eval scripts * Lua API for Redis Scripts * New Redis programibality page The PR also require changes on redis-io that was made [here](redis/redis-io#261)
yossigo
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, with one issue.
@MeirShpilraien Perhaps we should completely eliminate redis from the namespace when handling the load run? I think it might help prevent some confusion if we just expose a single function creation function.
|
@yossigo there are API's on |
|
we'll soon probably wanna add |
|
@MeirShpilraien I think we should opt-in on what we provide and not opt-out, and using a different name might make better distinction about these two run modes. |
|
@yossigo I see your point and I agree, so I guess we can introduce a new API, |
itamarhaber
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Haven't CRed, concept approved
The load run with only have 'library' API that contains only: * register_function * log
Those 2 commands add data to our dataset (we consider functions as data). So we want to disallow running those commands on OOM.
|
@yossigo @oranagra @itamarhaber @madolson |
madolson
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I reviewed the top level comment and read some of the code. I agree with all the decisions made.
Co-authored-by: Oran Agra <oran@redislabs.com>
1. rename 'library->redis' 2. better documentation for 'luaRegisterGlobalProtectionFunction'
|
@oranagra rename |
Co-authored-by: Oran Agra <oran@redislabs.com>
Redis Function Libraries
This PR implements Redis Functions Libraries as describe on: #9906.
Libraries purpose is to provide a better code sharing between functions by allowing to create multiple functions in a single command. Functions that were created together can safely share code between each other without worrying about compatibility issues and versioning.
Creating a new library is done using 'FUNCTION LOAD' command (full API is described below)
This PR introduces a new struct called libraryInfo, libraryInfo holds information about a library:
When Redis gets the
FUNCTION LOADcommand it creates a new empty libraryInfo. Redis passes theCODEto the relevant engine alongside the empty libraryInfo. As a result, the engine will create one or more functions by calling 'libraryCreateFunction'. The new funcion will be added to the newly created libraryInfo. So far Everything is happening locally on the libraryInfo so it is easy to abort the operation (in case of an error) by simply freeing the libraryInfo. After the library info is fully constructed we start the joining phase by which we will join the new library to the other libraries currently exist on Redis. The joining phase make sure there is no function collision and add the library to the librariesCtx (renamed from functionCtx). LibrariesCtx is used all around the code in the exact same way as functionCtx was used (with respect to RDB loading, replicatio, ...). The only difference is that apart from function dictionary (maps function name to functionInfo object), the librariesCtx contains also a libraries dictionary that maps library name to libraryInfo object.New API
FUNCTION LOAD
FUNCTION LOAD <ENGINE> <LIBRARY NAME> [REPLACE] [DESCRIPTION <DESCRIPTION>] <CODE>Create a new library with the given parameters:
Return "OK" on success, or error on the following cases:
Changed API
FUNCTION LIST
FUNCTION LIST [LIBRARYNAME <LIBRARY NAME PATTERN>] [WITHCODE]Command was modified to also allow getting libraries code (so
FUNCTION INFOcommand is no longer needed and removed). In addition the command gets an option argument,LIBRARYNAMEallows you to only get libraries that match the givenLIBRARYNAMEpattern. By default, it returns all libraries.INFO MEMORY
Added number of libraries to
INFO MEMORYCommands flags
DENYOOMflag was set onFUNCTION LOADandFUNCTION RESTORE. We consider those commands as commands that add new data to the dateset (functions are data) and so we want to disallows to run those commands on OOM.Removed API
Lua engine changes
When the Lua engine gets the code given on
FUNCTION LOADcommand, it immediately runs it, we call this run the loading run. Loading run is not a usual script run, it is not possible to invoke any Redis command from within the load run. Instead there is a new API provided bylibraryobject. The new API's:redis.log- behave the same asredis.logredis.register_function- register a new function to the libraryThe loading run purpose is to register functions using the new
redis.register_functionAPI. Any attempt to use any other API will result in an error. In addition, the load run is has a time limit of 500ms, error is raise on timeout and the entire operation is aborted.redis.register_functionredis.register_function(<function_name>, <callback>, [<description>])This new API allows users to register a new function that will be linked to the newly created library. This API can only be called during the load run (see definition above). Any attempt to use it outside of the load run will result in an error. The parameters pass to the API are:
Example
The following example creates a library called
libwith 2 functions,f1andf1, returns 1 and 2 respectively:Notice: Unlike
eval, functions inside a library get the KEYS and ARGV as arguments to the functions and not as global.Technical Details
On the load run we only want the user to be able to call a white list on API's. This way, in the future, if new API's will be added, the new API's will not be available to the load run unless specifically added to this white list. We put the while list on the
libraryobject and make sure thelibraryobject is only available to the load run by using lua_setfenv API. This API allows us to set theglobalsof a function (and all the function it creates). Before starting the load run we create a new fresh Lua table (call itg) that only contains thelibraryAPI (we make sure to set global protection on this table just like the general global protection already exists today), then we use lua_setfenv to setgas the global table of the load run. After the load run finished we updategmetatable and set__indexand__newindexfunctions to be_G(Lua default globals), we also pop out thelibraryobject as we do not need it anymore. This way, any function that was created on the load run (and will be invoke usingfcall) will see the default globals as it expected to see them and will not have thelibraryAPI anymore.An important outcome of this new approach is that now we can achieve a distinct global table for each library (it is not yet like that but it is very easy to achieve it now). In the future we can decide to remove global protection because global on different libraries will not collide or we can chose to give different API to different libraries base on some configuration or input.
Notice that this technique was meant to prevent errors and was not meant to prevent malicious user from exploit it. For example, the load run can still save the
libraryobject on some local variable and then using infcallcontext. To prevent such a malicious use, the C code also make sure it is running in the right context and if not raise an error.FUNCTION LISTpattern matching