aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
...
* Use #:duration-logger in datastore-call-with-transactionChristopher Baines2023-04-25
| | | | Rather than duplicating what it does.
* Split call-with-time-trackingChristopher Baines2023-04-25
| | | | So that you can observe a duration directly.
* Remove chunked input codeChristopher Baines2023-04-25
| | | | This is now upstream in Guile.
* Remove read-request-body workaroundChristopher Baines2023-04-25
| | | | | The Guile chunked input port now raises an exception when the input is incomplete.
* Log long database operationsChristopher Baines2023-04-25
| | | | | There's the delay logger that applies to all operations, but it would be helpful to see any long read or write as well.
* Support a duration logger in the worker thread channelChristopher Baines2023-04-25
|
* Increase the maximum number of reader threadsChristopher Baines2023-04-24
|
* Be rigerous about always using cached statementsChristopher Baines2023-04-24
| | | | | | The point here isn't really statement caching for performance, but using the cache of statements to keep track of them, and allowing them all to be finalized when the database connection is closed.
* Expire database connection threads on exceptionChristopher Baines2023-04-24
| | | | | | | | | | | I'm concerned that SQLite is very sensitive to prepared statements being left around without being finalized or reset, and I think this can happen if an exception occurs in a worker, but before the reset or finalize happens. Handling this in every individual case would be quite difficult, but given that guile-sqlite finalizes prepared statements when you close a connection, I think that mechanism can be used to somewhat rigerously avoid this (at least for cached prepared statements).
* Support expiring worker threads on exceptionChristopher Baines2023-04-24
|
* Add delay logging to more operations involved in submitting buildsChristopher Baines2023-04-24
|
* Log the build-id in datastore-find-build-derivation-systemChristopher Baines2023-04-24
| | | | As I'm not sure why a build wouldn't have a system.
* Reduce the extreme wal size thresholdChristopher Baines2023-04-24
| | | | To avoid the WAL getting too big.
* Avoid attempting to process build results twiceChristopher Baines2023-04-24
| | | | | The subsequent attempts fail due to the database constraints, but detecting this in the application is more elegant.
* Reverse the atomic-box-set! and vector-set! when storing eventsChristopher Baines2023-04-24
| | | | | | | | | | | | | Updating the current state id and index before writing the new entry to the vector is risky, since a thread could read the old entry if the timing is unlucky. Instead, write to the buffer first, then updating the current state id and index. This will cause problems for listeners reading from the very end of the buffer, but this can be a problem anyway if they're very behind, and they already handle that circumstance. Thanks to Ludo for reporting this!
* Set a custom printer for the build-coordinator recordChristopher Baines2023-04-24
| | | | In case this is causing issues with printing backtraces.
* Make sure the writer thread sets up before the reader threadsChristopher Baines2023-04-24
|
* Use an in memory table for the build_allocation_planChristopher Baines2023-04-23
| | | | Since this doesn't need to be persisted and changes often.
* Enable SQLite URI supportChristopher Baines2023-04-23
|
* Add delay logging for list related derivations with no buildChristopher Baines2023-04-23
|
* Fix hook processing error messageChristopher Baines2023-04-23
| | | | And rate limit the failures even if they occur in the hook processing code.
* Tweak the wal truncation againChristopher Baines2023-04-22
|
* Drop the wal checkpointing attemptsChristopher Baines2023-04-22
| | | | In case this helps with performance.
* Further tune the wal checkpointingChristopher Baines2023-04-22
| | | | | Try to be less disruptive, so allow the wal to grow a bit more, and wait less when the checkpointing is blocked.
* Remove some thread pool loggingChristopher Baines2023-04-22
| | | | As I'm not currently debugging this.
* Don't block writes for as long when attempting checkpointsChristopher Baines2023-04-22
|
* Tune the extreme WAL truncatingChristopher Baines2023-04-22
| | | | | Allow the WAL to grow a bit more, and increase the time to wait to truncate it.
* Try much harder to truncate the WALChristopher Baines2023-04-22
| | | | | | I think when there are lots of long reads taking place, that can prevent checkpointing and truncating the WAL, so allow waiting for much longer if this seems to be the case.
* Instrument the closing of writer threadsChristopher Baines2023-04-22
|
* Add some more exception handling around hooksChristopher Baines2023-04-22
|
* Tweak the busy timeouts for reading and writingChristopher Baines2023-04-22
| | | | | I'm not sure this will necessarily help with the database locking situation, but it's worth a try.
* Re-enable (backtrace) for the allocator threadChristopher Baines2023-04-22
| | | | This should be safe now that the <scheduler> record printer is in place.
* Don't use exceptions for blocked wal checkpointsChristopher Baines2023-04-22
| | | | As this clutters the output.
* Try to avoid problems with printing the fibers schedulerChristopher Baines2023-04-22
|
* Call sqlite step and reset/finalize together in most placesChristopher Baines2023-04-22
| | | | | Since calling reset or finalize is quite important, and this makes the code much simpler.
* Make datastore-call-with-transaction error messages more usefulChristopher Baines2023-04-21
|
* Fix agent-detailsChristopher Baines2023-04-21
|
* Deallocate canceled builds from agents when they startupChristopher Baines2023-04-21
|
* Include the submit_outputs information in the agent status responseChristopher Baines2023-04-21
| | | | | This means that agents will know whether to submit the outputs of builds, even if they're restarted.
* Include whether the build is canceled when listing agent buildsChristopher Baines2023-04-21
| | | | | | Canceled builds aren't unallocated from the agent immediately, so including the canceled status in the information given to the agent allows it to not start or potentially stop the build.
* Add some delay between optimize attemptsChristopher Baines2023-04-21
|
* Add some error handling to datastore-find-build-derivation-systemChristopher Baines2023-04-21
| | | | Not sure why I'm seeing errors here now, but I am.
* Add error handling around some date parsingChristopher Baines2023-04-21
|
* Handle COMMIT erroring with busy in datastore-call-with-transactionChristopher Baines2023-04-21
|
* Turns out COMMIT can return busy, so increase the timeoutChristopher Baines2023-04-21
|
* Add missing sqlite-step and sqlite-resetChristopher Baines2023-04-20
|
* Improve datastore-call-with-transactionChristopher Baines2023-04-20
| | | | | Include exception in log message, and separate out exception handling for the transaction start and body.
* Don't truncate the wal from reader threadsChristopher Baines2023-04-20
|
* Revert "Retry inside the worker thread"Christopher Baines2023-04-20
| | | | | | As this blocks other database operations that wouldn't wait. This reverts commit 42bf1086d588081548fb1efb0b4cf1e5ad119763.
* Drop the busy timeout for the write connectionChristopher Baines2023-04-20
| | | | Since I don't think it should be busy, so don't wait as long if it is.
generated by cgit v1.2.3 (git 2.45.0) at 2024-11-26 23:20:50 +0000