Badblocks

[OptimalDuck]

Hi,

I followed the steps in the resources to run badblocks and I'm curious if this is the expected behavior. I have two running at the same time and the bottom one is showing me a percentage done and seems normal. The first one however has a bunch of numbers just scrolling quickly. Is this what should be happening?

 

[WI_Hedgehog]

One option is to run multiple shells on the TrueNAS box....
Another is to run each job in the background with output piped to a different log file and use tail to watch file progress.

 

[OptimalDuck]

I believe I had them both running at the same time using tmux, but they aren't both behaving the same way, so it makes me think something has gone wrong.

 

[danb35]

The first one however has a bunch of numbers just scrolling quickly. Is this what should be happening?

No, it isn't. Those numbers are the block numbers that badblocks believes to be bad.

 

[OptimalDuck]

No, it isn't. Those numbers are the block numbers that badblocks believes to be bad.

That’s what I was starting to think. It is hard to see them all as they scroll, but it looks like it is listing every number. Is that likely telling me that something else is wrong? Or is it likely that all those blocks are bad on a new drive?

 

[WI_Hedgehog]

If badblocks returned a bunch of good blocks then a bunch of bad blocks...it's working and something went wrong. Did you do a SMART test first? What prompted you to use badblocks?

 

[OptimalDuck]

If badblocks returned a bunch of good blocks then a bunch of bad blocks...it's working and something went wrong. Did you do a SMART test first? What prompted you to use badblocks?

Its a new drive, so I was testing it before using it as suggested in the resource I linked to in the first post. I did run a SMART test first, and it didn't show any problems.

 

[joeschmuck]

Did you run a SMART Long/Extended test? Or just a Short test? If just a Short test, then run a Long test and see how that goes.

After you get done running badblocks, I'd check the SMART data again, then try badblocks again only on the suspect drive to verify the errors.

Sometimes you get a bad drive, hence the reason for testing them first. We hope all goes well but we all know that once in a blue moon something goes wrong.

 

[OptimalDuck]

I think i just ran short SMART tests. I missed that the guide had multiple tests listed.

 

[joeschmuck]

The short test does a test of the electronics and then a very quick surface test on a few areas, not the entire surface area like a Long test does. We recommend doing a SMART Long test first since it takes a fairly short period of time to run and if it finds problems then you do not need to use badblocks, which takes a long time to run since it runs 5 different test patterns.

We all learn something new.

 

[NugentS]

I just spent slightly less than two weeks testing 2*18TB disks with badblocks

 

[danb35]

There are a couple of reasons that badblocks is used in burn-in testing of the drives. Yes, there's the obvious of actually checking for bad blocks, but really the long SMART test ought to handle this. @OptimalDuck, I would suggest checking the SMART data for the drive in question (smartctl -a /dev/ada3, or whatever the drive number is); if there truly are bad blocks, they should show up there as well. But the other reason is simply to stress-test the drive for a while, to get it actively working for a while so that if it's going to fail, it will fail during testing rather than in production. It's not the only to do that, and it may not even be the best way to do that, but it's nonetheless effective.

 

[OptimalDuck]

Before I read everyone’s advice about the SMART tests I restarted badblocks and it seems to be going better this time. The first pass has completed and it’s onto the second (if I’m understanding it correctly). When it finishes I’ll run the SMART tests and see what happens. I think I may have messed things up while figuring out how to get it all setup the first time.

 

[joeschmuck]

If badblocks pass, all you should need to do is check the SMART attributes to see if all is good. I would also run a SMART Long test, only because that will be the test you will be running from this point forward to test your drives out. You should know that is passes without incident.

 

[OptimalDuck]

If badblocks pass, all you should need to do is check the SMART attributes to see if all is good. I would also run a SMART Long test, only because that will be the test you will be running from this point forward to test your drives out. You should know that is passes without incident.

How long should a long SMART test typically take on a 4tb drive?

 

[Redcoat]

How long should a long SMART test typically take on a 4tb drive?

One of mine just advised 452 minutes.

 

[OptimalDuck]

My badblocks completed without errors and I ran the long SMART tests and they also had no errors so I appear to be good. I'm not sure what I did wrong the first time, but it seems like it was a me problem. Thanks for the help everyone.

 

[joeschmuck]

I'm not sure what I did wrong the first time

That doesn't matter, the end result is the drives are good right now. Glad it worked out for you.

 

[Etorix]

How long should a long SMART test typically take on a 4tb drive?

smarctl -a /dev/$DRIVE tells you how long the test is (where $DRIVE is to be replaced by the appropriate da#, ada#, sdX).