Date: Thu, 5 Jul 2018 08:09:07 +0700 From: Eugene Grosbein <eugen@grosbein.net> To: George Mitchell <george+freebsd@m5p.com>, freebsd-hackers@freebsd.org Subject: Re: Confusing smartd messages Message-ID: <5B3D6FB3.5010908@grosbein.net> In-Reply-To: <17c1cb78-f2f0-ff13-6425-8ac634ddaa56@m5p.com> References: <dfccd275-954c-11da-1790-e75878f89ad1@m5p.com> <5B3D6975.2060508@grosbein.net> <17c1cb78-f2f0-ff13-6425-8ac634ddaa56@m5p.com>
next in thread | previous in thread | raw e-mail | index | archive | help
05.07.2018 7:54, George Mitchell write: >> First step is running zpool scrub or even replace the disk and run "dd if=/dev/zero of=/dev/ada1". >> [...] > > I'll start with zpool scrub. What's the magic command (if any) to > force a rewrite of the sectors? -- George Modern HDDs are embedded computers theyselves with their own CPU, RAM, flash memory and special operating system (firmware) running inside. They have spare load of generally unused and hidden sectors also. When firmware sees write error due to really bad spot on magnetic part of disk making a sector "bad", it performs a remap by means of internal translation table for LBAs to use one of hidden sectors instead of bad one. And until the amount of hidden sectors is not exhausted (there are can be over a thousand of them), the disk keeps running nearly fine. Sometimes the sector just gets wrong checksum written due to power failure, hence read errors due to "mismatched" checksum. Overwrite corrects such "soft bad" error without need to remap. Eugene P.S. your smtp server (MX) rejects my direct mail to you with "550 5.7.1 Access denied" errors.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5B3D6FB3.5010908>