How to Troubleshoot Regular Backup Failures?

[Eruanantion]

Hello everybody. I'm a new home-labber (been doing this less than a year) but I've quickly become obsessed. I have a 2-node PVE setup and I've hooked up a PBS as well. This has been going well for a while until I actually needed to restore one of my backups from PBS this week and I realized that roughly 80% of my backups are failing verification with messages similar to this: Data blob has wrong CRC checksum. My professional background is in software, so I'm learning most of the hardware stuff as I go. Some AI-assisted digging leads me to believe that I most likely have a faulty disk (I'm using a 5-year-old 2TB external spinner for my backups). But I wanted to ask the pros before I start making changes.

What's the best way to figure out how to fix this? In PBS the disk is showing that it's passed the SMART tests, so I'm not sure how to proceed.

 

[Eduardo Taboada]

It sounds a problem with the underlaying disks of storage. Have you tried a SMART check on disks?

 

[Onslow]

Welcome to the Forum, @Eruanantion .
Unless your RAM is of ECC type it can also be the culprit.
You can also try verifying the backups again. I have seen situations where next verification succeeds.

 

[Eruanantion]

It sounds a problem with the underlaying disks of storage. Have you tried a SMART check on disks?

I have, and it passed all of the tests. I can paste the output here if you think it'll help.

Welcome to the Forum, @Eruanantion .
Unless your RAM is of ECC type it can also be the culprit.
You can also try verifying the backups again. I have seen situations where next verification succeeds.

It's DDR3 RAM. I did consider that old RAM might be the issue since my PBS is an 11-year-old laptop (and I happen to know it was having RAM issues when I replaced it as my main computer), but everything I dug up made it sound like the disk would be much more likely to be the culprit.

I think when I have a chance, I may buy another disk to plug in and see if that fixes my issues.

 

[Eduardo Taboada]

I have, and it passed all of the tests. I can paste the output here if you think it'll help.


It's DDR3 RAM. I did consider that old RAM might be the issue since my PBS is an 11-year-old laptop (and I happen to know it was having RAM issues when I replaced it as my main computer), but everything I dug up made it sound like the disk would be much more likely to be the culprit.

I think when I have a chance, I may buy another disk to plug in and see if that fixes my issues.

Sometimes RAM errors in fact can be due a power supply failure. RAM needs precise voltage to reliably store data; too low voltage causes data corruption and may be the cause of the problem also.