Morgan Tocker
Morgan Tocker
When spirit runs, there are 3 common log lines that keep being printed: ``` 2024-07-08T23:51:17.639 info finished periodic flush of binary log: total-duration=2.854617ms batch-size=5583 2024-07-08T23:51:47.636 info migration status: state=copyRows copy-progress=273732111/3306757995...
Currently if there is a checksum failure, or the process is killed during checksum it requires a complete restart. Since checksums typically take about 10% of copier time, we might...
This looks to have been recently introduced: ``` 2024-06-27T20:07:49.4809983Z --- FAIL: TestMetadataLockContextCancel (0.00s) 2024-06-27T20:07:49.4810995Z panic: runtime error: invalid memory address or nil pointer dereference [recovered] 2024-06-27T20:07:49.4812203Z panic: runtime error: invalid...
``` 2024-06-28T14:08:43.1734561Z runner_test.go:2666: 2024-06-28T14:08:43.1735065Z Error Trace: /home/runner/work/spirit/spirit/pkg/migration/runner_test.go:2666 2024-06-28T14:08:43.1735657Z /opt/hostedtoolcache/go/1.22.4/x64/src/runtime/asm_amd64.s:1695 2024-06-28T14:08:43.1736084Z Error: Error "sql: database is closed" does not contain "context canceled" 2024-06-28T14:08:43.1736417Z Test: TestResumeFromCheckpointE2EWithManualSentinel 2024-06-28T14:08:43.1737236Z time="2024-06-28T14:08:21Z" level=info msg="Starting spirit migration:...
We use sysbench to test [Spirit](https://github.com/cashapp/spirit) (an online schema change tool). Spirit has an optimization where it can [ignore keys that have been modified above a certain known point](https://github.com/cashapp/spirit?tab=readme-ov-file#ignore-key-above-watermark). Unfortunately...
I discussed this with @tirsen. I would like to disable vendoring and later (not this PR) bump the go version.
Forked from review comment in https://github.com/cashapp/spirit/pull/389: - Currently we have a goroutine monitor a channel for table changes. - Since it can't return an error, we should allow that goroutine...
## A Pull Request should be associated with an Issue. > We wish to have discussions in Issues. A single issue may be targeted by multiple PRs. > If you're...
We have seen case of mdl locks leaking. Manual inspection of the code doesn't show an obvious case of what causes it, because the `defer cancel()` in runner.Run() should clean...
We saw a scenario where a long running schema change was canceled and then started again one week later. There were still binary logs available, but it was a significant...