You could be there a long time...

What in particular makes it more prone to fail in the boot process? I thought the wear would mostly be an issue due to swap file usage, and I'd have thought usage of that is minimal until some RAM hungry apps are open?

Perhaps you could speed up the failure by reducing the system RAM, and Perhaps having the OS launch something on startup that will wear the disk a bit, before it reboots?

Heh nice thread, I'm really interested to see where this goes.
I assume your goal is to simulate a normal usage scenario and reach a conclusion along the lines of "it survived 1000 boot processes and if you factor in the actual usage during these times, this sd probably won't last a year" (or the opposite)?

RetroPCCupboard wrote on Yesterday, 14:42:

You could be there a long time...

What in particular makes it more prone to fail in the boot process? I thought the wear would mostly be an issue due to swap file usage, and I'd have thought usage of that is minimal until some RAM hungry apps are open?

Perhaps you could speed up the failure by reducing the system RAM, and Perhaps having the OS launch something on startup that will wear the disk a bit, before it reboots?

XP writes to the registry hives, event log, & NTFS journal ($LogFile) on every reboot, so without wear leveling, those spots will get worn. It also writes to the page file too.

Reducing the ram would increase the use of the page file, but since it's usually cheap and easy and satisfying to add more ram to your retro XP computer, why would you run a XP system without it?

Reboots arent what murders SDcards used this way.

Abusive numbers of tiny writes all over the place does.

Browser caches, temp files from installers, and idiotic/braindead pagefile operations.

THOSE kill SDCards.

[I used a microSD as /home on a hacked chromebook running xubuntu. Firefox was THE major contributor to the first go-around of trying this, as was the syslog, but browser cache was the primary offender. It took about 3 months to die from being used as a daily driver, before I got wise, and instituted mitigations. After doing that, the whole chromebook failed before the microSD did, with several YEARS of service given. Web browser cache is by far and away the most abrasive thing to an SDCard used this way.]

konc wrote on Yesterday, 15:42:

Heh nice thread, I'm really interested to see where this goes.
I assume your goal is to simulate a normal usage scenario and reach a conclusion along the lines of "it survived 1000 boot processes and if you factor in the actual usage during these times, this sd probably won't last a year" (or the opposite)?

Some people fret about running solid state storage without wear leveling or trim support, for well known reasons.

I just want to put some numbers around it to gage how worried you should be. 16GB SD with NTFS is probably a worst case scenario.

If it dies after 250 reboots, I think we can all agree that this isn't how you want to build an XP system.

If it makes it to 2500 reboots, then maybe it's not the biggest factor to worry about if you are willing to buy an new SD every 5 years.

wierd_w wrote on Yesterday, 19:29:

Reboots arent what murders SDcards used this way. […]
Show full quote

Reboots arent what murders SDcards used this way.

Abusive numbers of tiny writes all over the place does.

Browser caches, temp files from installers, and idiotic/braindead pagefile operations.

THOSE kill SDCards.

Browser cache files? Who uses a web browser on XP on a daily basis in 2026?

Temp files from installers? How often do you install software on an XP build?

Page file? A legit concern, but it will get worked pretty heavily during the reboot process.

I'm not gonna pretend that some bespoke business internal webapp that needs unsafe activeX and IE6 does not still haunt the world in some way, and that some poor schmuck is not left holding the bag looking for an inexpensive alternative to spinning rust. 😁

But for 'ordinary use', like playing vintage vidya gaems, absolutely. SDCard is plenty fast enough (on par with ATA 33 speeds, even at the worst), and should work fine, as long as pagefile is kept off the darn thing.

OK, the great rebootinating has begun!!!

1. 16GB SD card allocated and validated with ValiDrive
2. XP is installed & registered
3. one batchfile added to the startup folder that kicks off a vbscript
4. one VB script created that logs the current reboot count to a text file and then reboots after 20 seconds

Each reboot cycle is taking a second or two less than 1 minute

douglar wrote on Yesterday, 21:28:

OK, the great rebootinating has begun!!! […]
Show full quote

OK, the great rebootinating has begun!!!

1. 16GB SD card allocated and validated with ValiDrive
2. XP is installed & registered
3. one batchfile added to the startup folder that kicks off a vbscript
4. one VB script created that logs the current reboot count to a text file and then reboots after 20 seconds

Each reboot cycle is taking a second or two less than 1 minute

Just checking... is the text file being written to a different drive? Would be a shame to lose all the data if\when the card dies.

I am interested to hear how long it lasts though!

I'm hoping that the card is still somewhat readable after failing. If it isn't, I'll put the reboot log on a separate drive.

452 reboots after 7 hours, still going

douglar wrote on Today, 04:16:

I'm hoping that the card is still somewhat readable after failing. If it isn't, I'll put the reboot log on a separate drive.

452 reboots after 7 hours, still going

If it were me, the log would be going on a different drive. If the card lasts way longer than you expect and then the file can't be recovered it would be a bit maddening.

I'd just pop in a flash drive and save the log to that.