Filter by reset
- Everywhere
1789 - Issues
176 - Pull Requests
400 - Comments
1213
Autosharding tests: Outage & Rebalancing
Validate rebalancing during node outages while maintaining the replication factor. Ignore rebalancing when any part of the shards is missing. Ensure we are not performing a rebalance during a short-term node outage. Check shard rebalancing when a new node …
est::size_L
Auto-sharding rollback on fail
Proposal: Currently, as per MVP, we do not have a rollback mechanism. We should think about and introduce a recovery system so that in case something goes wrong in creating a table or the rebalancing mechanism, we can rollback all intermediate tables/ …
Auto-sharding polling mechanism
Proposal: As discussed before, currently we rely on the queue system ONLY, but due to the blocking nature of queries like alter cluster add table ... , it's not guaranteed that even if it fails and returns an error, the cluster is not synced. Instead of …
Wrong Content-Type for /metrics
Bug Description: The /metrics endpoint returns an incorrect Content-Type header. Example response: ~  curl -is 0:19308/metrics|head -10 HTTP/1.1 200 OK Server: 10.1.0 c63706038@25060608 (columnar 5.0.1 59c7092@25060304) (secondary 5.0.1 59c7092@25060304) …
bug
Unstable fuzzy results with default layouts after disk chunk flush
Bug Description: The same select returns an empty result after flush ramchunk . Example: mysql -P9306 -h0 -v -e "drop table if exists t; create table t(name text) min_infix_len='2' diskchunk_flush_write_timeout = '9999'; insert into t values(1, 'LIMITED …
bug
need live tests for queries to 'information schema.tables' table
Proposal: Files in src/Plugin/Select implement responses to different tools speaking with mysql proto. Since some funcs going to move to daemon's code, query tests are necessary. That is all real kinds of 'select ... from information_schema.tables' - with …
Clearly describe what each plugin does
Proposal: Now that our plugins are growing in size and complexity, we need to improve their documentation. It's currently quite difficult to understand what each plugin does and which SQL/HTTP commands it expects. I suggest we add a README file to each …
Buddy loose records in log (debugv mode)
Bug Description: #!/usr/bin/env bash set -x docker stop kafka && docker rm kafka docker stop manticore && docker rm manticore wget -O /tmp/import.sh https://raw.githubusercontent.com/manticoresoftware/manticoresearch-buddy/main/test/Kafka/import.sh wget - …
bugest::size_S
Support for REPLACE in auto schema
Bug Description: We need to support auto table creation for REPLACE queries just as we do for INSERT ones. E.g., the following query must create the test table if it does not exist: REPLACE INTO test(f) VALUES(1) Manticore Search Version: dev Operating …
bugest::size_M
Applying specific table settings when creating a table by autoschema
Proposal: Sometimes, a table auto-created by a request from an Elastic tool (e.g., Kibana) requires specific settings to be applied ( e.g., enabling min_infix_len , etc.). We need to be able to recognize such requests by their endpoints and apply required …
est::size_S
Proposal: We should add logic that will support our own bulk ( /bulk ) endpoint and parse it the same way we do for /_bulk to make insertion into distributed tables possible. As discussed before, there are several things to check: Whether it works with …
We already have the option to skip loading the selected plugin when the daemon starts, but if we try to skip installing many plugins, any further conditions will be skipped when we go beyond the number of characters that can be processed. Reproducing this …
bug
Proposal: Currently, we support Apache Superset 3.x, but version 4.x is now available on the market. We should validate our existing support and implement any missing handlers to fully support the new version. Checklist: To be completed by the assignee. …
est::NO_ESTIMATE
Confirm that we cannot alter or change the replicated cluster from the user side (only Buddy).
We need to ensure that changes to the cluster, necessary for replicating shards in the auto-sharding scheme, can't be made through direct requests but are only permitted from the user side. This task is on hold due to: https://github.com/manticoresoftware …
We should add support for altering sharded tables. Currently, there is no alter supported on the distributed table, so we can catch it with Buddy and process the logic. Altering a sharded table should drop the distributed table, alter all sharded tables, …
enhancementest::size_M
Verify that we cannot alter or change the sharded table from the user side (only Buddy)
When we create a sharded table with Buddy, we must ensure that users are not allowed to directly alter, modify, or delete this table. These operations should be performed exclusively by Buddy because the table is distributed and contains multiple shards. …
Rename all tables we use in Queue plugin to be started with _
Proposal: As we discussed earlier, we should rename our system-created tables in the queue plugin to start with an underscore. They should also be prefixed with the plugin name, so all tables should follow the format queue[something]. This fulfills our …
est::size_S
Autosharding tests: Perf & heavy load
Measure and validate latency from query execution to final results on different shards and replication factors. Ensure that we have enough timeout for waiting for the blocking "create table" command to return a response to the user in case of high shard …
est::size_L
The following request returns a json response instead of a formatted message, when repeated: curl 0:9308/cli -d "create table products(title text, price float) morphology='stem_en'" Query OK, 0 rows affected (0.002 sec) curl 0:9308/cli -d "create table …
Buddy plugins don't work from other plugin
I'm trying to create a new plugin. I am adding this closure: $taskFn = static function (Payload $handler, HTTPClient $manticoreClient): TaskResult { $result= $manticoreClient->sendRequest('show queries')->getResult(); /** @var array{error?:string} $result …
waiting