bugfix on service discovery, and rework pf the dashboard

This commit is contained in:
Pierre
2026-02-02 00:02:55 +01:00
parent 9b4a802ff3
commit 7886870046
13 changed files with 923 additions and 707 deletions
@@ -103,4 +103,7 @@ export interface ServiceInfo {
/** Last heartbeat timestamp (ISO string) */
lastHeartbeat: string;
/** Role for meta-core instances: 'leader', 'follower', or undefined for other services */
role?: string;
}
+2 -3
View File
@@ -80,15 +80,14 @@ export class KVManager {
logger.info(`Starting KVManager for ${this.config.serviceName}...`);
// Initialize service discovery (meta-core handles actual registration)
// We keep this for service discovery (reading other services)
// Initialize and start service discovery for registration and heartbeat
const apiUrl = this.config.baseUrl || `http://${hostname()}:${this.config.apiPort}`;
this.serviceDiscovery = new ServiceDiscovery({
metaCorePath: this.config.metaCorePath,
serviceName: this.config.serviceName,
apiUrl
});
// Note: We don't start ServiceDiscovery - meta-core handles registration
await this.serviceDiscovery.start();
// If direct Redis URL provided, skip leader discovery
if (this.config.redisUrl) {
+81 -25
View File
@@ -42,18 +42,87 @@ export class LeaderClient {
private watcher: FSWatcher | null = null;
private onChangeCallbacks: (() => void)[] = [];
// URL caching
private cachedUrls: URLsResponse | null = null;
private urlsCacheTime: number = 0;
private readonly urlsCacheTTL: number = 5000; // 5 seconds
constructor(config: LeaderClientConfig) {
this.config = config;
this.infoFilePath = `${config.metaCorePath}/locks/kv-leader.info`;
}
/**
* Read leader info from the info file
* Read API URL from file (plain text format)
*/
private async getApiUrlFromFile(): Promise<string | null> {
try {
const content = await fs.readFile(this.infoFilePath, 'utf-8');
return content.trim() || null;
} catch (error: any) {
logger.error(`Failed to read API URL from file: ${error.message}`);
return null;
}
}
/**
* Fetch URLs from meta-core /urls API with caching
*/
private async fetchUrls(apiUrl: string): Promise<URLsResponse | null> {
// Check cache
const now = Date.now();
if (this.cachedUrls && (now - this.urlsCacheTime) < this.urlsCacheTTL) {
return this.cachedUrls;
}
try {
const response = await fetch(`${apiUrl}/urls`, {
method: 'GET',
headers: { 'Accept': 'application/json' },
signal: AbortSignal.timeout(5000)
});
if (!response.ok) {
logger.error(`Failed to fetch URLs: ${response.status} ${response.statusText}`);
return null;
}
this.cachedUrls = await response.json() as URLsResponse;
this.urlsCacheTime = now;
return this.cachedUrls;
} catch (error: any) {
logger.error(`Error calling /urls API: ${error.message}`);
return null;
}
}
/**
* Read leader info from file and /urls API
*/
async getLeaderInfo(): Promise<LeaderLockInfo | null> {
try {
const content = await fs.readFile(this.infoFilePath, 'utf-8');
this.leaderInfo = JSON.parse(content) as LeaderLockInfo;
// Read API URL from file (plain text)
const apiUrl = await this.getApiUrlFromFile();
if (!apiUrl) {
return null;
}
// Fetch full info from /urls API
const urls = await this.fetchUrls(apiUrl);
if (!urls) {
return null;
}
// Convert URLsResponse to LeaderLockInfo
this.leaderInfo = {
hostname: urls.hostname,
baseUrl: urls.baseUrl,
apiUrl: urls.apiUrl,
redisUrl: urls.redisUrl,
webdavUrl: urls.webdavUrl,
timestamp: Date.now(),
pid: 0 // Unknown for remote leader
};
return this.leaderInfo;
} catch (error: any) {
logger.error(`Failed to read leader info: ${error.message}`);
@@ -91,12 +160,11 @@ export class LeaderClient {
*/
async getUrls(): Promise<URLsResponse | null> {
// First try using configured metaCoreUrl
let apiUrl = this.config.metaCoreUrl;
let apiUrl: string | null = this.config.metaCoreUrl ?? null;
// Fall back to reading from leader info
// Fall back to reading from file
if (!apiUrl) {
const info = await this.getLeaderInfo();
apiUrl = info?.apiUrl;
apiUrl = await this.getApiUrlFromFile();
}
if (!apiUrl) {
@@ -104,23 +172,7 @@ export class LeaderClient {
return null;
}
try {
const response = await fetch(`${apiUrl}/urls`, {
method: 'GET',
headers: { 'Accept': 'application/json' },
signal: AbortSignal.timeout(5000)
});
if (!response.ok) {
logger.error(`Failed to get URLs: ${response.status} ${response.statusText}`);
return null;
}
return await response.json() as URLsResponse;
} catch (error: any) {
logger.error(`Error calling /urls API: ${error.message}`);
return null;
}
return this.fetchUrls(apiUrl);
}
/**
@@ -157,7 +209,11 @@ export class LeaderClient {
try {
this.watcher = watch(lockDir, (eventType, filename) => {
if (filename === 'kv-leader.info') {
logger.debug(`Lock file ${eventType}, checking for leader...`);
logger.debug(`Lock file ${eventType}, invalidating cache...`);
// Invalidate cache to force fresh API call
this.cachedUrls = null;
this.urlsCacheTime = 0;
this.getLeaderInfo().then(() => {
this.notifyChange();
}).catch((err) => logger.error(`Error reloading leader info: ${err.message}`));
@@ -10,10 +10,26 @@
import { Redis } from 'ioredis';
import { Logger } from 'tslog';
import * as os from 'os';
import type { IKVClient, KeyValuePair } from './IKVClient.js';
const logger = new Logger({ name: 'RedisClient' });
/**
* Stream message from Redis Streams
*/
export interface StreamMessage {
id: string;
type: 'batch' | 'reset' | 'plugin:complete';
payload: string;
timestamp: string;
}
/**
* Stream consumer callback
*/
export type StreamMessageHandler = (message: StreamMessage) => Promise<void>;
/**
* Full metadata structure including fields needed for virtual path computation
*/
@@ -72,6 +88,11 @@ export class RedisClient implements Partial<IKVClient> {
private reconnectTimer: NodeJS.Timeout | null = null;
private subscriptions: Map<string, (message: string) => void> = new Map();
// Stream consumer state
private streamConsumerRunning = false;
private streamConsumerAbort: AbortController | null = null;
private consumerName: string;
constructor(config: RedisClientConfig = {}) {
this.config = {
url: config.url ?? process.env.REDIS_URL ?? 'redis://127.0.0.1:6379',
@@ -79,6 +100,9 @@ export class RedisClient implements Partial<IKVClient> {
filesVolume: config.filesVolume ?? process.env.FILES_VOLUME ?? '/files',
reconnectInterval: config.reconnectInterval ?? 5000,
};
// Unique consumer name for this instance
this.consumerName = `${os.hostname()}-${process.pid}`;
}
/**
@@ -163,6 +187,9 @@ export class RedisClient implements Partial<IKVClient> {
* Disconnect from Redis
*/
async disconnect(): Promise<void> {
// Stop stream consumer first
this.stopStreamConsumer();
if (this.reconnectTimer) {
clearTimeout(this.reconnectTimer);
this.reconnectTimer = null;
@@ -442,6 +469,224 @@ export class RedisClient implements Partial<IKVClient> {
}
}
// ========================================================================
// Redis Streams Consumer Methods
// ========================================================================
/**
* Initialize stream consumer group
* Creates the consumer group at position 0 to read all historical events
*
* @param stream - Stream name (e.g., 'meta-sort:events')
* @param group - Consumer group name (e.g., 'meta-fuse-vfs')
*/
async initStreamConsumer(stream: string, group: string): Promise<void> {
if (!this.client) {
throw new Error('Redis not connected');
}
try {
// Create consumer group at position 0 (read all historical events)
// MKSTREAM creates the stream if it doesn't exist
await this.client.xgroup('CREATE', stream, group, '0', 'MKSTREAM');
logger.info(`Created consumer group '${group}' for stream '${stream}'`);
} catch (error: any) {
// BUSYGROUP means group already exists - that's fine
if (error.message?.includes('BUSYGROUP')) {
logger.debug(`Consumer group '${group}' already exists`);
} else {
throw error;
}
}
}
/**
* Process pending entries from crashed consumers
* Uses XAUTOCLAIM to claim entries that have been idle too long
*
* @param stream - Stream name
* @param group - Consumer group name
* @param minIdleTime - Minimum idle time in ms before claiming (default: 30000)
* @param onMessage - Handler for each message
*/
async processPendingEntries(
stream: string,
group: string,
minIdleTime: number = 30000,
onMessage: StreamMessageHandler
): Promise<number> {
if (!this.client) {
return 0;
}
let processed = 0;
let cursor = '0-0';
try {
while (true) {
// XAUTOCLAIM claims idle entries and returns them
const result = await this.client.xautoclaim(
stream,
group,
this.consumerName,
minIdleTime,
cursor,
'COUNT',
100
) as [string, Array<[string, string[]]>, string[]];
const [nextCursor, entries] = result;
cursor = nextCursor;
if (!entries || entries.length === 0) {
break;
}
for (const [id, fields] of entries) {
try {
const message = this.parseStreamEntry(id, fields);
if (message) {
await onMessage(message);
// ACK the message after successful processing
await this.client.xack(stream, group, id);
processed++;
}
} catch (error: any) {
logger.error(`Error processing pending entry ${id}:`, error.message);
}
}
// If cursor is 0-0, we've processed all pending entries
if (cursor === '0-0') {
break;
}
}
if (processed > 0) {
logger.info(`Processed ${processed} pending stream entries`);
}
} catch (error: any) {
logger.error('Error processing pending entries:', error.message);
}
return processed;
}
/**
* Start stream consumer loop
* Reads messages using XREADGROUP and calls the handler
*
* @param stream - Stream name
* @param group - Consumer group name
* @param onMessage - Handler for each message
* @param blockMs - Block timeout in ms (default: 5000)
*/
async startStreamConsumer(
stream: string,
group: string,
onMessage: StreamMessageHandler,
blockMs: number = 5000
): Promise<void> {
if (!this.client) {
throw new Error('Redis not connected');
}
if (this.streamConsumerRunning) {
logger.warn('Stream consumer already running');
return;
}
this.streamConsumerRunning = true;
this.streamConsumerAbort = new AbortController();
logger.info(`Starting stream consumer: ${stream} (group: ${group}, consumer: ${this.consumerName})`);
// Consumer loop
while (this.streamConsumerRunning && !this.streamConsumerAbort.signal.aborted) {
try {
// XREADGROUP blocks waiting for new messages
// '>' means only read new messages (not already delivered to this consumer)
// Use call() to bypass TypeScript's strict argument checking for xreadgroup
const result = await (this.client.call(
'XREADGROUP',
'GROUP', group,
this.consumerName,
'BLOCK', blockMs,
'COUNT', 10,
'STREAMS', stream,
'>'
) as Promise<Array<[string, Array<[string, string[]]>]> | null>);
if (!result || result.length === 0) {
continue; // Timeout, loop again
}
// Process each stream's messages
for (const [streamName, entries] of result) {
for (const [id, fields] of entries) {
try {
const message = this.parseStreamEntry(id, fields);
if (message) {
await onMessage(message);
// ACK the message after successful processing
await this.client!.xack(stream, group, id);
}
} catch (error: any) {
logger.error(`Error processing stream entry ${id}:`, error.message);
// Don't ACK on error - message will be reprocessed
}
}
}
} catch (error: any) {
if (this.streamConsumerAbort?.signal.aborted) {
break;
}
logger.error('Stream consumer error:', error.message);
// Brief pause before retrying
await new Promise(resolve => setTimeout(resolve, 1000));
}
}
logger.info('Stream consumer stopped');
}
/**
* Stop the stream consumer loop
*/
stopStreamConsumer(): void {
if (!this.streamConsumerRunning) {
return;
}
logger.info('Stopping stream consumer...');
this.streamConsumerRunning = false;
this.streamConsumerAbort?.abort();
this.streamConsumerAbort = null;
}
/**
* Parse a stream entry into a StreamMessage
*/
private parseStreamEntry(id: string, fields: string[]): StreamMessage | null {
// Fields come as flat array: ['type', 'batch', 'payload', '...', 'timestamp', '123']
const fieldMap: Record<string, string> = {};
for (let i = 0; i < fields.length; i += 2) {
fieldMap[fields[i]] = fields[i + 1];
}
if (!fieldMap.type || !fieldMap.payload) {
logger.warn(`Invalid stream entry ${id}: missing type or payload`);
return null;
}
return {
id,
type: fieldMap.type as StreamMessage['type'],
payload: fieldMap.payload,
timestamp: fieldMap.timestamp || '0',
};
}
/**
* Get stats about Redis data
*/
@@ -7,8 +7,8 @@
* Services can discover each other by reading these JSON files.
* Stale detection: if lastHeartbeat > 60 seconds ago, service is considered stale.
*
* Note: meta-core (Go sidecar) handles actual service registration.
* This class is primarily used for discovering other services.
* Each service must call start() to register itself and begin heartbeat updates.
* meta-core runs a cleanup process that removes services with stale heartbeats.
*/
import { promises as fs } from 'fs';
@@ -79,9 +79,8 @@ export class ServiceDiscovery {
}
/**
* Register this service
* Note: In the new architecture, meta-core handles registration.
* This is kept for backwards compatibility and testing.
* Register this service by writing service info to the shared filesystem.
* meta-core's cleanup process monitors these files and removes stale entries.
*/
async register(): Promise<void> {
await this.ensureServicesDir();
@@ -292,8 +291,7 @@ export class ServiceDiscovery {
// ========================================================================
/**
* Full startup sequence
* Note: In the new architecture, meta-core handles registration.
* Full startup sequence: register service and start heartbeat loop.
*/
async start(): Promise<void> {
if (this.isStarted) return;