beobal commented on code in PR #4630: URL: https://github.com/apache/cassandra/pull/4630#discussion_r2904014619
########## src/java/org/apache/cassandra/db/SystemPeersValidator.java: ########## @@ -0,0 +1,227 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.cassandra.db; + +import java.net.InetAddress; +import java.util.ArrayList; +import java.util.Collections; +import java.util.HashMap; +import java.util.HashSet; +import java.util.List; +import java.util.Map; +import java.util.Objects; +import java.util.Set; + +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import org.apache.cassandra.cql3.UntypedResultSet; +import org.apache.cassandra.db.marshal.UTF8Type; +import org.apache.cassandra.db.virtual.PeersTable; +import org.apache.cassandra.locator.InetAddressAndPort; +import org.apache.cassandra.schema.SchemaConstants; +import org.apache.cassandra.tcm.ClusterMetadata; +import org.apache.cassandra.tcm.membership.Location; +import org.apache.cassandra.tcm.membership.NodeAddresses; +import org.apache.cassandra.tcm.membership.NodeId; +import org.apache.cassandra.utils.FBUtilities; + +import static org.apache.cassandra.cql3.QueryProcessor.executeInternal; +import static org.apache.cassandra.db.SystemKeyspace.LEGACY_PEERS; +import static org.apache.cassandra.db.SystemKeyspace.PEERS_V2; + +/** + * Validator to ensure system.peers and system.peers_v2 tables match ClusterMetadata on startup. + * This is critical for backward compatibility as older clients and tools read from these + * legacy tables while TCM uses ClusterMetadata as the source of truth. + * + * The validator detects inconsistencies and automatically repairs them by synchronizing + * the peers tables with the current ClusterMetadata. + */ +public class SystemPeersValidator +{ + private static final Logger logger = LoggerFactory.getLogger(SystemPeersValidator.class); + + public static void validateAndRepair(ClusterMetadata metadata) + { + Map<InetAddressAndPort, UntypedResultSet.Row> peersV2Rows = getPeersV2Rows(); + Map<InetAddress, UntypedResultSet.Row> legacyPeersRows = getLegacyPeersRows(); + + Map<InetAddressAndPort, NodeId> expectedEndpoints = new HashMap<>(); + Map<InetAddress, NodeId> expectedAddresses = new HashMap<>(); + for (NodeId nodeId : getExpectedPeerNodes(metadata)) + { + InetAddressAndPort endpoint = metadata.directory.endpoint(nodeId); + expectedEndpoints.put(endpoint, nodeId); + expectedAddresses.put(endpoint.getAddress(), nodeId); + } + + String deleteV2Query = String.format("DELETE FROM %s.%s WHERE peer = ? AND peer_port = ?", + SchemaConstants.SYSTEM_KEYSPACE_NAME, PEERS_V2); + for (InetAddressAndPort endpoint : peersV2Rows.keySet()) + { + if (!expectedEndpoints.containsKey(endpoint)) + { + logger.info("Removing stale peer {} from {}", endpoint, PEERS_V2); + executeInternal(deleteV2Query, endpoint.getAddress(), endpoint.getPort()); + } + } + + String deleteLegacyQuery = String.format("DELETE FROM %s.%s WHERE peer = ?", + SchemaConstants.SYSTEM_KEYSPACE_NAME, + LEGACY_PEERS); + for (InetAddress address : legacyPeersRows.keySet()) + { + if (!expectedAddresses.containsKey(address)) + { + logger.info("Removing stale peer {} from {}", address, LEGACY_PEERS); + executeInternal(deleteLegacyQuery, address); + } + } + + for (Map.Entry<InetAddressAndPort, NodeId> entry : expectedEndpoints.entrySet()) + { + NodeId nodeId = entry.getValue(); + InetAddressAndPort endpoint = entry.getKey(); + UntypedResultSet.Row v2Row = peersV2Rows.get(endpoint); + UntypedResultSet.Row legacyRow = legacyPeersRows.get(endpoint.getAddress()); + + List<String> v2Discrepancies = collectV2Discrepancies(v2Row, nodeId, metadata); Review Comment: > It tells an operator exactly what's wrong — e.g. Updating peer 10.0.0.1 in peers_v2 for stale fields [data_center, rack] yes, IMO this is unnecessary and given the frequency with which I would anticipate this actually happening, I don't think it's worth the code bloat. > With Row::toString(), we will get pipe-delimited positional values without column names, and the operator would have to figure out the diff manually to identify what's stale. I think this is fine, in the unlikely situation that a mismatch is detected having the full row logged would be good enough, figuring out the diff manually would pretty straightforward. > I did consider it, but reverted that change because I found it more convenient to see the complete set of validations for a given table at a glance without jumping between methods. What I had in mind was something like: ``` private static boolean peersV2RowIsEquivalent(UntypedResultSet.Row row, NodeId nodeId, ClusterMetadata metadata) { NodeAddresses addresses = metadata.directory.getNodeAddresses(nodeId); return commonColumnsAreEquivalent(row, nodeId, addresses, metadata) && row.has("preferred_port") && Objects.equals(row.getInetAddress("preferred_port"), addresses.broadcastAddress.getPort()) && row.has("rpc_port") && Objects.equals(row.getInetAddress("rpc_port"), addresses.nativeAddress.getPort()); } private static boolean peersRowIsEquivalent(UntypedResultSet.Row row, NodeId nodeId, ClusterMetadata metadata) { NodeAddresses addresses = metadata.directory.getNodeAddresses(nodeId); return commonColumnsAreEquivalent(row, nodeId, addresses, metadata); } private static boolean commonColumnsAreEquivalent(UntypedResultSet.Row row, NodeId nodeId, NodeAddresses addresses, ClusterMetadata metadata) { if (row == null) return false; Location location = metadata.directory.location(nodeId); // This column is differently named in the peers and peers_v2 tables String nativeAddressColumn = row.has("native_address") ? "native_address" : "rpc_address"; // Check existence first because row.getXXX can NPE if the column is not present return row.has("preferred_ip") && Objects.equals(row.getInetAddress("preferred_ip"), addresses.broadcastAddress.getAddress()) && row.has(nativeAddressColumn) && Objects.equals(row.getInetAddress(nativeAddressColumn), addresses.nativeAddress.getAddress()) && row.has("data_center") && Objects.equals(row.getString("data_center"), location.datacenter) && row.has("rack") && Objects.equals(row.getString("rack"), location.rack) && row.has("host_id") && Objects.equals(row.getUUID("host_id"), nodeId.toUUID()) && row.has("release_version") && Objects.equals(row.getString("release_version"), metadata.directory.version(nodeId)) && row.has("schema_version") && Objects.equals(row.getUUID("schema_version"), metadata.schema.getVersion()) && row.has("tokens") && Objects.equals(row.getSet("tokens", UTF8Type.instance), tokensAsSet(metadata.tokenMap.tokens(nodeId))); } ``` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
