atlasdb
atlasdb copied to clipboard
palantir
Transactional Distributed Database Layer
Endpoints need to be updated to reflect the new Conjure APIs e.g. bleh:8421/timelock/api/tl/ts/myNamespace instead of bleh:8421/timelock/api/myNamespace/timestamp/fresh-timestamp
As the name suggests: this currently isn't handled.
From: https://github.com/palantir/atlasdb/blob/develop/atlasdb-client/build.gradle#L77 I do not believe that atlasdb-client should have a compile dependency on rocksdb. At minimum, it should be a light dependency which doesn't require downstream consumers to pull...
As discussed with @jeremyk-91 and @j-baker
This line can potentially throw if there are connection issues with the lock server. https://github.com/palantir/atlasdb/blob/962e597b5af6a976cba990aee6e3108f02b52ebd/atlasdb-impl-shared/src/main/java/com/palantir/atlasdb/transaction/impl/SnapshotTransactionManager.java#L251
Suppose I have an unowned Executor that I pass to a LockServiceImpl. This will spawn two tasks, the token and grant reaper, that will _rename_ LockServiceImpl threads. However, when the...
Since 0.236.8 or before, we are failing internal testing with the following error: ```java.lang.IllegalStateException: The classes below are Enums with non-final fields or non-unique @UniqueValue fields: com.palantir.atlasdb.keyvalue.cassandra.pool.CassandraClientPoolHostLevelMetric```
See PDS-120860. Suppose a timelock cluster has 3 nodes, and node 1 was the leader but then got partitioned off. If node 1 doesn't get any more requests, it could...
We have cluster management docs for upgrading a 1 to 3 node cluster. These use curls from before the conjure java migration, so something like ``` curl -XGET https://:/timelock/api/leader//timelock/fresh-timestamp ```...