Time Machine Disconnects Due to RAM Saturation by restsdk-server (My Cloud EX2 Ultra, OS 5) — UI Control Fails

[Herold]

Subject: CRITICAL: Time Machine Disconnects Due to RAM Saturation by restsdk-server (My Cloud EX2 Ultra, OS 5) — UI Control Fails

Hello WD Community and Support Team,I am reporting a critical resource conflict issue on my WD My Cloud EX2 Ultra (Firmware 5.31.108 and running My Cloud OS 5) that causes repeated Time Machine ™ backup failures. I have identified the root cause using the Command Line Interface (CLI) and found a working—but highly abnormal—workaround.

---

Problem Description

Time Machine backups from my macOS system 26.1 (Tahoe) consistently fail after the preparation phase, resulting in the error message: “Disk disconnected., backup failed”

This failure started occurring after my NAS filled up with a significant amount of data (several terabytes), suggesting a resource scaling issue with the mandatory indexing service.

---

CLI Diagnostic Findings (Root Cause)

Analysis via SSH showed a severe memory contention:

• RAM Saturation: The NAS’s memory (approx. 1GB RAM) was completely saturated.

• Culprit Processes (high %VSZ / %MEM): The restsdk-server (WD’s indexing/Cloud API process) was using over 90% of the RAM, and the otaclient (Over-The-Air update client) was also reserving large memory blocks.

• The Conflict: The massive RAM reservation by these services caused system timeouts during the heavy I/O required by Time Machine, leading to the forced disconnection.

Crucially (UI Control Fails): We confirmed that the problem is not resolvable via the GUI:

• I attempted to disable the service by switching Share Access > Mobile & Web App Access > OFF on all shares and then rebooting the NAS.

• Despite this, the restsdk-server immediately relaunched after the reboot, consuming over 90% of the NAS’s RAM. This confirms that the indexing service is considered mandatory and cannot be disabled through the provided GUI options.

Resolution (Abnormal Workaround)

Since the WD watchdog daemon immediately relaunches these processes, the only successful method to run a stable backup is to manually suspend them via SSH before initiating Time Machine:

1. Stop/Suspend Processes: Before starting a backup, I must manually freeze the memory-hogging processes:

   Bash

   # Check top for current PIDs first, then execute:
   kill -STOP [PID_restsdk-server]
   kill -STOP [PID_otaclient]
   
   

2. Run Backup: The Time Machine backup then runs to completion successfully.

3. Resume Processes: After the backup finishes, I must manually resume the processes:

   Bash

   kill -CONT [PID_restsdk-server]
   kill -CONT [PID_otaclient]
   
   

---

Request to WD Support

Since this workaround is required for a fundamental NAS function (backup) and points to a major resource management flaw in My Cloud OS 5, I have two questions:

1. Permanent Fix: Since the UI controls (including Share Access > Mobile & Web App Access > OFF) are ineffective, are there any official, non-CLI solutions to permanently disable or limit the resource usage of the restsdk-server and otaclient processes?

2. Future Updates: Is the WD team aware of this crippling RAM saturation issue on the My Cloud OS 5 platform, and are there plans to release a firmware update that properly throttles these services during high-load I/O operations (like Time Machine)?

Thank you for your assistance.