What causes "Delete Batch Stream On Error: ${e.message}" in qishibo/AnotherRedisDesktopManager?

Reconnect and restart the batch delete; already-found keys remain listed so the retry is cheaper

qishibo/AnotherRedisDesktopManager · error

Delete Batch Stream On Error: ${e.message}

Error message

Delete Batch Stream On Error: ${e.message}

What it means

The SCAN-based readable stream used by the batch-delete dialog emitted its 'error' event. Mid-scan failures come from the underlying connection or cluster: ECONNRESET, READONLY during failover, MOVED/IO errors when slots move in cluster mode, or the server closing the stream. The handler stops loadingScan and shows a 1.5s toast, but the already-scanned keys remain in the list.

Source

Thrown at src/components/DeleteBatch.vue:95

          count: 20000,
        };

        const stream = node.scanBufferStream(scanOption);
        this.scanStreams.push(stream);

        stream.on('data', (keys) => {
          this.addToList(keys.sort());

          // pause for dom rendering
          stream.pause();
          setTimeout(() => {
            this.loadingScan && stream.resume();
          }, 100);
        });

        stream.on('error', (e) => {
          this.loadingScan = false;
          this.$message.error({
            message: `Delete Batch Stream On Error: ${e.message}`,
            duration: 1500,
          });
        });

        stream.on('end', () => {
          // all nodes scan finished(cusor back to 0)
          if (--this.scanningCount <= 0) {
            this.loadingScan = false;
            this.scanningEnd = true;
          }
        });
      });
    },
    addToList(keys) {
      const list = [];
      for (const key of keys) {
        list.push({ key, str: this.$util.bufToString(key) });

View on GitHub (pinned to c149855106)

Solutions

Reconnect and restart the batch delete; already-found keys remain listed so the retry is cheaper
Run long scans against a stable topology (no concurrent resharding/failover)
Tune SCAN COUNT and the pause interval so iterations finish before idle timeouts
Persist the cursor and resume from it instead of restarting from 0

Example fix

// before
stream.on('error', (e) => {
  this.loadingScan = false;
  this.$message.error({ message: `Delete Batch Stream On Error: ${e.message}`, duration: 1500 });
});

// after
stream.on('error', (e) => {
  this.loadingScan = false;
  stream.destroy();
  this.$message.error({
    message: `Delete Batch Stream On Error: ${e.message} (scanned ${this.allKeysList.length} keys, retry to continue)`,
    duration: 3000,
  });
});

Defensive patterns

Strategy: try-catch

Try / catch

stream.on('error', (e) => {
  this.loadingScan = false;
  stream.destroy(); // release cursor and socket bookkeeping
  // keep already-collected keys so a retry can continue
  this.$message.error(`Delete Batch Stream On Error: ${e.message}`);
});

Prevention

Always listen for 'error' on readable streams or the process throws
Destroy the stream in the error handler
Track scanned keys/cursor so retries resume instead of restarting
Avoid long scans during cluster maintenance windows

When it happens

Trigger: Scanning a huge keyspace when the connection drops; cluster resharding or master failover mid-scan; server/client idle timeout killing long SCAN iterations; LB cutting long-lived connections.

Common situations: Deleting millions of keys over VPN, cleanup jobs running during maintenance windows, scans outliving load-balancer idle timeouts.

Related errors

AI-assisted analysis of qishibo/AnotherRedisDesktopManager@c149855106 (2026-08-22). Data as JSON: /api/errors/a5eb883193955676. Report an issue: GitHub.