atlasdb
atlasdb copied to clipboard
palantir
Transactional Distributed Database Layer
PDS-186437 Scrubber is slow and cannot be sped up so hard deleting many things is not really feasible whereas targeted sweep can be scaled. Another issue is that if conservative...
Follow up from https://github.com/palantir/atlasdb/pull/5517 / https://github.com/palantir/atlasdb/pull/5535 to migrate the places that `@SuppressWarnings("ProxyNonConstantType")` to use [proxy processor](https://github.com/palantir/proxy-processor) to enable atlasdb usage from graal native image.
See internal link: https://g.p.b/foundry/mp/pull/12089#discussion_r1462867 AtlasDB has a number of components which need to be initialized (one of which is the transaction manager). If the initialization does not succeed on the...
@jkozlowski let me know that we apparently still depend on http2 error-handling (!), for some funky legacy compatibility with old http-remoting clients when attempting to acquire the Backup Lock (!!)...
**Goals (and why)**: **Implementation Description (bullets)**: **Testing (What was existing testing like? What have you done to improve it?)**: **Concerns (what feedback would you like?)**: **Where should we start reviewing?**:...
In particular, it's quite useful to know when number of active connections is trending towards/pinned at max as a way to alert on or diagnose connection leaks. These metrics might...
Currently, `ConnectionConfig#getHikariConfig` bootstraps a MetricsRegistry via `SharedMetricRegistries#getOrCreate` (see [.here](https://github.com/palantir/atlasdb/blob/249de1a6f3e21eb3907451eb232d96066f659194/atlasdb-dbkvs-hikari/src/main/java/com/palantir/nexus/db/pool/config/ConnectionConfig.java#L143) However, some users may have their own `MetricRegistry` that they wish to bootstrap into the config. We can use `HikariConfig#setMetricRegistry` after...
May be useful for some users. If you say run a task with an expected number of reads, we can error if the user tries to do another read even...
Also provide some doc that this retriable version should not be used for long running read only transactions. T runTaskReadOnlyWithRetry(TransactionTask task) throws E; `lockImmutableTsOnReadOnlyTransactions = true` may cause read only...
