Constantly have to run `xfs_repair` on one drive

[joelrfernandes]

Hello all,

 

Honestly I'm not a big expert on the hardware side of things but, I'm finding myself having to run `xfs_repair` on one particular drive over and over again. The symptoms are always the same: eventually some directories stop providing file lists, or one Docker container starts having issues with a directory, etc. When that happens I always have to run `xfs_repair -vL /dev/md8` to get that drive back again. I'm attaching the output of one of the last times I ran xfs_repair. I'm also attaching the diagnostics log.

 

Any help would really be appreciated.

monster-diagnostics-20171118-2219 (1).zip

xfs_repair.log

[irishdab]

Your drive ST4000DM000-1F2168_S30091TE-20171118-2219 in the smart data is listed as failing now. Is that the drive you keep having to run xfs_repair on?

Edit to be more clear the smart data shows the error is:

 

184 End-to-End_Error        0x0032   094   094   099    Old_age   Always   FAILING_NOW 6

https://kb.acronis.com/content/9119

Best to replace that disk in my view.

 

[joelrfernandes]

No, that's not the drive. In fact, that notice has been there for a while now (maybe over a year) but I never had issues with that drive so far. I do remember searching here at the time and I think I read that error might not cause any harm, but yes, I should still replace it.

 

The drive that I constantly run xfs_repair is disk8

[joelrfernandes]

Now reading the description of that attribute, I should really replace that drive soon.

[JorgeB]

3 hours ago, joelrfernandes said:

Now reading the description of that attribute, I should really replace that drive soon.

 

You should, as for disk8 if the filesystem keeps getting corrupt you should move all data to another disk, reformat and move data back.

[SSD]

It is certainly in the.manufacture's best interest to not report a disk failing when it is not. So when they say it is failing, I tend to believe it.

 

But failing the end to end attribute is very rare without other symptoms. And you say you've used for over a year without it getting worse. 

 

My general rule of thumb is if you can run three consecutive parity checks and the smart attributes don't change, that the drive may have 'stabilized' (whatever that means).

 

But I would replace it and use this drive for backups if it seems pretty reliable. Or better yet rma if under warranty.

[joelrfernandes]

SSD: Unfortunately it's not under warranty anymore  But yes, I'll definitely replace it and use it for something else.

 

As for disk8, I'm going to follow your advice, johnnie.black. I'll get an extra drive, move contents there, re-format it and move data back to it.