Re: Known and unfixed active data loss bug in MM + XFS with large folios since Dec 2021 (any kernel from 6.1 upwards)

[Christian Theune]

> On 11. Oct 2024, at 09:27, Christian Theune <ct@xxxxxxxxxxxxxxx> wrote:
> 
> I’m going to gather a few more instances during the day and will post them as a batch later.

I’ve received 8 alerts in the last hours and managed to get detailed, repeated walker output from two of them:

- FC-41287.log
- FC-41289.log

The other logs are tracebacks as the kernel reported them but the situation resolved itself faster than I could log in and run the walker script. In FC-41289.log I’m also providing output from `ps auxf` to see what the process tree looks like, maybe that helps, too.

My observations: 

- different entry points from the XFS code: unlink, f(data)sync, truncate
- in none of the cases I caught I could see any real competing traffic (aside from maybe occasional journal writes and very little background noise), all affected machines are staging environments that saw basically no usage during that timeframe

I’m stopping my alerting now as it’s been interrupting me every few minutes and I’m running out of steam sitting around waiting for the alert. ;)

Christian

Attachment: FC-41281.log
Description: Binary data

Attachment: FC-41282.log
Description: Binary data

Attachment: FC-41283.log
Description: Binary data

Attachment: FC-41285.log
Description: Binary data

Attachment: FC-41286.log
Description: Binary data

Attachment: FC-41287.log
Description: Binary data

Attachment: FC-41288.log
Description: Binary data

Attachment: FC-41289.log
Description: Binary data

-- 
Christian Theune · ct@xxxxxxxxxxxxxxx · +49 345 219401 0
Flying Circus Internet Operations GmbH · https://flyingcircus.io
Leipziger Str. 70/71 · 06108 Halle (Saale) · Deutschland
HR Stendal HRB 21169 · Geschäftsführer: Christian Theune, Christian Zagrodnick