Essential Insights on Windows PCs You Might Not Know
Written on
One crucial aspect of using Windows PCs is often overlooked by users: the inevitable failure of your computer. It’s a tough reality that can lead to a panic when you realize all your data might be lost. The distress of waiting for a tech service to determine if your files are recoverable can be overwhelming.
While you may be familiar with Windows Backup features, this article aims to provide insights that could be both familiar and new to you. Unless you've worked in data recovery or IT, the information here is likely to prove beneficial.
Take it from my experience: I have cherished photos stored solely on an old iPod, which currently functions, but the data is increasingly inaccessible due to outdated parts.
> If you rely on your PC for personal or professional use, it's crucial to understand how to safeguard your data from potential loss and how to recover it—especially if it includes irreplaceable family memories, as I discovered when my previous Windows PC failed.
The most prevalent causes of PC failure include faulty memory and hard drive (HDD/SSD) issues. This is particularly concerning for work computers and is frequently encountered with consumer-grade hardware.
Consumer laptops, especially the slim models, are particularly vulnerable to failures under typical usage conditions. Desktop computers, conversely, tend to handle heat and performance better.
When your computer fails, you face two pressing questions: - How soon can you resume your work? - Is it possible to recover any of your lost data?
To address both, you must navigate the complex domain of Bare Metal Recovery (BMR), which simply means recovering data onto different hardware after a failure.
In larger enterprises, it’s common to have minimal downtime, as virtual machines (VMs) can be activated while hardware repairs are underway.
For tech-savvy individuals, this process may be manageable. However, for the majority who prefer not to delve into technical details or invest in advanced solutions, creating an image backup is the most efficient method for safeguarding and restoring your entire system.
Although image backups should ideally work as seamlessly as Mac backups, the reality for Windows users is often far from it.
There are two significant challenges with Windows image backups: - Successfully creating reliable image backups over time can be quite tricky. - Restoring from a complete image backup can be even more difficult.
Given these challenges, it's advisable to choose a reputable backup service like Acronis:
https://www.acronis.com/en-us/products/true-image/features/imaging/
While I’m not endorsing this product, I personally use it, and it generally performs better than Windows Backup. My experience with recovery, however, was disappointing. I found the recovery options confusing and ultimately unusable.
- I will update this if I discover a more effective tool.
Despite Following Best Practices, I Faced Disappointments
In light of my experiences, I've opted for RAID 5 configurations on my work PCs for several years, even on my work laptop, which accommodates three NVMe SSDs.
Initially, I didn’t require much storage on my main office desktop, so I set up a RAID 5 with three 250GB Samsung 970 Evo NVMe SSDs. When one drive failed and I needed additional space, I expanded my setup to include four SSDs in a RAID 5 array.
How to Prevent Catastrophic Failures on Your Work PC with RAID 5
Recently, I noticed a decline in performance over the past six months, which gradually worsened.
After expanding my RAID 5 setup, I discovered that Intel VROC has a bug affecting RAID 5 configurations with more than three SSDs. This led to slow performance and repeated corruption in my development environment.
I mistakenly set the stripe size to 32KB, which was too low for my needs. After analyzing my files, I found that the average file size for 434k files was 427.29KB, indicating the stripe size was causing issues.
I eventually settled on a maximum stripe size of 128KB, as that's the largest supported by Intel.
> Note: When files are distributed across disks in a RAID configuration, it can lead to complications with development tools that store states in files larger than the maximum stripe size.
To resolve Windows corruptions, you should run:
- These tools often identify and repair issues, even if your PC seems to be functioning well.
Action Plan & Strategy
My Visual Studio was becoming unusable due to daily corruption. I identified the following potential culprits:
- Too small a stripe size (32KB)
- Known issue with Intel VROC RAID 5 causing parity errors with more than three drives
Despite opening a support ticket with Intel, they did not acknowledge this as a known issue. Users remain uncertain whether it’s a reporting glitch or a genuine bug, but I can confirm the corruption.
- Intel RAID 5 Write Hole Issue leading to corruption
This issue was supposedly resolved with a setting I hadn’t previously used, but its effectiveness remains uncertain.
My plan involved taking a full image backup, deleting, and recreating the RAID 5 with a maximum stripe size around 157KB. If corruption persisted, I would have to replace three drives with larger SSDs for both space and reliability.
Here’s what transpired:
- I performed a Windows 10 image backup, which ultimately succeeded, but the process was incredibly tedious.
This was executed via Control Panel > Backup & Restore (Windows 7). Many online resources detail how to set this up. I allowed Windows to manage the backup settings, particularly ensuring I kept only the most recent system image to avoid issues with accumulating backups.
For those struggling to create the first backup, consider using an SSD initially, then transferring it to an HDD for subsequent incremental backups.
Following a reboot, I had to run chkdsk, which identified and fixed numerous issues before I could successfully take an image backup.
- I created a Dell Recovery Drive using a Dell utility, which proved ineffective with an image backup. Although it allows for factory resets, it doesn't grant access to essential admin functions.
I also created a Windows 10 Recovery Drive, which ultimately worked, albeit with considerable difficulty.
To prepare for the worst, I backed up crucial files to a USB drive in case I needed to reinstall everything if the image restore failed.
After taking every precaution, I deleted the existing RAID array and recreated it with a larger stripe size. When working with Intel VROC, avoid enabling options that span RAID across multiple Intel VMD Controllers, as this restricts the array to data drives only.
Upon attempting to restore from the image backup, I encountered a series of failures:
- OS Recovery did not succeed
- Dell Recovery couldn't restore from backup
- Acronis failed to recognize the image during recovery attempts
I spent an entire night and much of the following day troubleshooting strange errors.
How I Rescued My Work PC
From past experiences, I learned that image backups are sensitive to fitting perfectly within the available disk space during restoration. If they don't align, you'll encounter cryptic errors.
The best backup companies offer a customized Windows Recovery Environment (Win RE) to ensure the backup image can fit into the target disk space. The boundaries of the backup image can be adjusted to match the available space, provided it’s at least as large as the used space in the backup image.
I realized that even though I was setting up the RAID on the same SSDs, the available free space was impacted by the change in stripe size from 32KB to 128KB. Additionally, my prior mistake of “expanding” the RAID disk during an OS drive extension had consequences.
https://www.easeus.com/partition-manager/epm-pro.html
This program allows you to rearrange partitions on your disk, positioning free space next to your existing drive, making it easier to extend your primary drive.
Using this utility, I ultimately recreated the RAID 5 with three 500GB Samsung 970 Evo Plus NVMe SSDs, which allowed the restore process to proceed smoothly, even with the previously ineffective OS recovery drive.
I repurposed my old 970 Evo Plus as a journaling drive to mitigate the RAID Write Hole issue from Intel, as not utilizing it would have been a missed opportunity.
- Closing the RAID Write Hole disables disk caching, which significantly impacts performance.
- After consulting with Dell Sales, I acquired a 1000VA APC UPS.
- I had to disable the RAID Write Hole to enable disk caching in Windows and the RAID utility.
- With the UPS providing about 21 minutes of backup power, I feel secure with the write hole closed, as I no longer anticipate unscheduled power outages.
Everything seems to be functioning well, and I will provide updates as I continue using my work PC.
In the meantime, I've prepared additional 64GB USB drives for future recovery needs.
Summary
To effectively avert complete disaster, consider these essential measures:
USB 3.0 Drive (> 20GB) with Windows Recovery Disk
This process can take some time and is tailored to your specific PC.
USB 3.0 Drive (> 20GB) with Manufacturer-Specific Recovery Disk
This is particularly useful for Dell users, as it includes hardware-specific drivers.
USB 3.0 Drive (> 20GB) with Windows 10 Installation Media
Utilize the second button to create reusable Windows 10 installation media.
A Complete Image Backup of Your PC
These are fundamental safeguards for anyone using PCs for business or personal purposes, especially for preserving family photos and documents.
Many opt to store backup images remotely, whether in the cloud or at a different physical location, to mitigate disaster risks.
- This is highly advisable for small businesses.
- I would certainly recommend this for family photos and videos.
March 22, 2021
I acquired additional USB drives today:
https://www.amazon.com/gp/product/B081QSHG13/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1
Creating a Windows 10 recovery disk is essential, even without prior backups. This ensures you can recover your Windows installation, even if it requires a fresh start.
To create the recovery drive, search for "recovery" in Windows, follow the prompts, and ensure you select options to back up system files.
- The process can take some time, so patience is key.
To be continued...