Tor Brede Vekterli
Tor Brede Vekterli
When using multiple content cluster groups in self-hosted Vespa there is a requirement to specify a `distribution` attribute that states how replicas are to be distributed across groups (see https://docs.vespa.ai/en/performance/sizing-examples.html)....
Our current backend metric aggregation implementations are built around explicit snapshotting every _N_ minutes, where each such snapshot effectively resets the tracked metric value internally. In other words, a counter,...
Cluster controller ZooKeeper writes should be gated with a test-and-set on expected zknode version
Even though the cluster controller has a leader election algorithm built on top of a strongly consistent store (ZooKeeper), leaders today blindly write new state to ZooKeeper. This means that...
Distributor should log and emit metrics when it is not able to send bucket info requests to a node
When a cluster state changes, a distributor will send bucket info requests to all nodes that are marked as available in the cluster state. These are all nodes that have...
@geirst please review schema/C++ code (primary focus on `routable_factories_8.cpp`), @jonmv please review schema/Java code (primary focus on `RoutableFactories80.java` and its codec design). Note that this commit by itself will **fail**...
### Background Vespa uses algorithmic data placement, which means all cluster nodes need to agree on the inputs to the algorithm to be able to compute the same output. This...
If a Vespa C++ process terminates unexpectedly (i.e. from raising `SIGSEGV`, `SIGABRT` etc.) and core dumps are disabled on the host, very little information is left for debugging the root...
@geirst please review. Most added lines are tests. @baldersheim FYI Avoids potentially having to deserialize the entire update just to get to a single bit of information that is technically...
@hakonhall please review. Once merged, we should observe the output of this in practice for a bit before wiring it to anything. With these changes the cluster controller continuously maintains...