I have a physical HPE server running Windows Server 2019 with iLO 5 v2.44. It starts normally, but applications begin crashing one after another until the system becomes unusable. RDP reports "An internal error has occurred," and the remote console eventually goes black. SystemPropertiesRemote.exe reports that its side-by-side configuration is incorrect, while SideBySide Event 59 mentions invalid XML syntax in the ndfapi.dll manifest. There are also repeated DCOM Event 10000 errors involving DllHost.exe and "Only part of a ReadProcessMemory request was completed" messages. The graphical interface crashes before DISM or SFC can complete. iLO does not show obvious hardware errors beyond network-related entries. What should I test first to determine whether this is failing hardware, storage or RAID corruption, memory, or a damaged Windows installation?
5 Answers
If external diagnostics are stable but the Windows files remain badly corrupted, repairing or rebuilding the installation may be faster than chasing every damaged component. Before reinstalling, make sure backups are valid and identify the underlying cause so the replacement installation is not corrupted again.
The symptoms could easily come from storage corruption rather than several unrelated application failures. Check the RAID controller and physical disks, review controller logs, and consider whether a power interruption occurred while write-back cache was unavailable. Also verify the controller and drive firmware. A clean iLO log does not rule out disk, controller, or memory faults.
Run a proper memory stress test, such as an extended MemTest86 pass, and check temperatures and hardware diagnostics. Bad RAM can corrupt files and cause apparently random crashes, invalid manifests, and incomplete memory-read errors. Do not assume the problem is only Windows until memory and storage have been tested.
If the hardware appears healthy, boot into Windows Recovery or WinPE and run offline disk checks. First use DiskPart to confirm the actual Windows and system-partition drive letters, then run CHKDSK against the correct volume. DISM and SFC can also be run offline, but a recovery environment may need the HPE RAID driver loaded manually with drvload before it can see the array.
Start by separating hardware problems from Windows problems. Boot the server from a known-good WinPE or Linux environment, then leave it running and perform basic disk and memory checks. If the live environment also crashes or shows I/O errors, focus on RAM, overheating, the RAID controller, firmware, or the disks. If it stays stable, the Windows installation or its storage volume is more likely damaged.
Safe Mode or a recovery console is also worth trying if the system cannot stay up long enough to boot external media.