Bitcoin Forum

What a relief. It took a week of fiddling to rescue a bit of BTC that was stuck in an Armory wallet on Windows. Bitcoin Core has destroyed two of my hard drives (i.e. a Seagate Barracuda and a SanDisk Ultra SSD). A Samsung EVO SSD was able to withstand the beating. I was getting sporadic "Checksum mismatch, data corrupted" and "Deserialize or I/O error" fatal errors after long download periods and had to reindex/restart the blockchain sync many times which was very painful. Posting in case it helps someone trying understand why they get weird corruption errors during a sync.

It's something else going wrong with your hardware. I have been running core on old drives and new drives on fast drives and slow drives. Never had one die.
Sounds like there is something else marginal in your setup and the bit of stress that core causes when it's syncing is causing issues.
My guess is RAM.

-Dave

That's news.
I know that synchronization is I/O extensive, but HDD/SSD shouldn't fail that easily unless there's something wrong with the hardware.

Have you set the dbcache (database cache) to an insane value that your RAM can't handle together with the OS's background processes?

It could really just be the drives. Sometimes they just fail.
Another possibility would be the SATA cable.

Indeed, this would definitely be the case if the drives had a poor SMART health test results to begin with.

Note: never run Core on a hard disk with failing SMART check results, even if it's just a degree or two above normal temperature.

I know that some series of Seagate Barracuda hard drives showed high percentage of failures but it's not only related to Bitcoin Core and it happened even to regular users who used it for storing of data and surfing.
Important thing for hard drives is to do regular health checks from time to time, scan for errors and use trim option for SSD drives.
If same failure thing happened with your SSD drive than maybe you have issue with some different hardware component of your computer.

Intensive I/O activity won't make HDD suddenly fail, but only reduce it's lifespan (especially if you mine chia coin). I use 3-4 years old to store Bitcoin Core data without any problem.


It's unlikely, but another possibility is buggy chipset on motherboard. For example, Intel Sandy Bridge's chipset had bug on SATA port (https://www.cnet.com/news/intels-sandy-bridge-chipset-flaw-the-fallout/ (https://www.cnet.com/news/intels-sandy-bridge-chipset-flaw-the-fallout/)).

Most (all?) systems will stop at POST / Boot when your drive is giving SMART errors with a warning and a press a key to continue message.



That was why I was thinking RAM, that the drives are not dying but bad RAM is causing corruption.
Most other things will show other issues in other places. That or something else on the I/O chain.


I would think that at that point even the 3rd drive would be giving errors / have issues.

-Dave
 

Some of you guys may remember that few months ago I had some issues with my hard drive that was randomly disconnecting and even shutting down my computer.
After I scanned it for errors it showed some unrecoverable error and bad health, and I was already preparing that I have to buy a new drive,
but all issues were gone after I replaced cables, changed port and double checked connections.
I don't know if this is the case of faulty cables, but it's always worth checking it out.