Return an error when scheduling a reducer with a long delay #77
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Prior to this commit, it was possible for a module to crash SpacetimeDB by scheduling a reducer with a delay longer than ~2yrs. This was due to our use of `tokio_utils::time::DelayQueue` to handle scheduling. `DelayQueue`'s internal data structure imposes a limit of 64^6 ms on delays, a little more than two years. Attempting to insert with a delay longer than that panics. With this commit, we avoid the panic by checking ourselves that the requested delay is not longer than 64^6 ms. This requires bubbling a `ScheduleError` up from `Scheduler::schedule` to `WasmInstanceEnv::schedule`, where it is converted into a `RuntimeError` which crashes the module. `Scheduler::schedule` could also fail because its transaction to compute a new id was fallible. This seems unlikely to ever fail, and if it does, we have bigger problems, so `unwrap`ping might still be reasonable for that case, but this commit converts it into a handle-able `Err`or anyway, as there's essentially no cost in complexity to doing so.
jdetter
approved these changes
Jul 13, 2023
There was a problem hiding this comment.
Tested, SpacetimeDB no longer crashes. We don't really get a meaningful error in the case of using spacetime call from the command line:
Error: Response text: The Wasm instance encountered a fatal error.
Caused by:
HTTP status server error (530 <unknown status code>) for url (http://localhost:3000/database/call/93dda09db9a56d8fa6c024d843e805d8/test)
And also it appears that the module logs are empty. I think this error is going to be extremely uncommon so this is fine for now.
cloutiertyler
pushed a commit
that referenced
this pull request
Aug 1, 2023
Prior to this commit, it was possible for a module to crash SpacetimeDB by scheduling a reducer with a delay longer than ~2yrs. This was due to our use of `tokio_utils::time::DelayQueue` to handle scheduling. `DelayQueue`'s internal data structure imposes a limit of 64^6 ms on delays, a little more than two years. Attempting to insert with a delay longer than that panics. With this commit, we avoid the panic by checking ourselves that the requested delay is not longer than 64^6 ms. This requires bubbling a `ScheduleError` up from `Scheduler::schedule` to `WasmInstanceEnv::schedule`, where it is converted into a `RuntimeError` which crashes the module. `Scheduler::schedule` could also fail because its transaction to compute a new id was fallible. This seems unlikely to ever fail, and if it does, we have bigger problems, so `unwrap`ping might still be reasonable for that case, but this commit converts it into a handle-able `Err`or anyway, as there's essentially no cost in complexity to doing so.
cloutiertyler
pushed a commit
that referenced
this pull request
Aug 1, 2023
Prior to this commit, it was possible for a module to crash SpacetimeDB by scheduling a reducer with a delay longer than ~2yrs. This was due to our use of `tokio_utils::time::DelayQueue` to handle scheduling. `DelayQueue`'s internal data structure imposes a limit of 64^6 ms on delays, a little more than two years. Attempting to insert with a delay longer than that panics. With this commit, we avoid the panic by checking ourselves that the requested delay is not longer than 64^6 ms. This requires bubbling a `ScheduleError` up from `Scheduler::schedule` to `WasmInstanceEnv::schedule`, where it is converted into a `RuntimeError` which crashes the module. `Scheduler::schedule` could also fail because its transaction to compute a new id was fallible. This seems unlikely to ever fail, and if it does, we have bigger problems, so `unwrap`ping might still be reasonable for that case, but this commit converts it into a handle-able `Err`or anyway, as there's essentially no cost in complexity to doing so.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description of Changes
Prior to this commit, it was possible for a module to crash SpacetimeDB by scheduling a reducer with a delay longer than ~2yrs. This was due to our use of
tokio_utils::time::DelayQueueto handle scheduling.DelayQueue's internal data structure imposes a limit of 64^6 ms on delays, a little more than two years.Attempting to insert with a delay longer than that panics.
With this commit, we avoid the panic by checking ourselves that the requested delay is not longer than 64^6 ms.
This requires bubbling a
ScheduleErrorup fromScheduler::scheduletoWasmInstanceEnv::schedule, where it is converted into aRuntimeErrorwhich crashes the module.Scheduler::schedulecould also fail because its transaction to compute a new id was fallible. This seems unlikely to ever fail, and if it does, we have bigger problems, sounwrapping might still be reasonable for that case, but this commit converts it into a handle-ableError anyway, as there's essentially no cost in complexity to doing so.API
If the API is breaking, please state below what will break