Re: Disk failure
Dan Lukes <[email protected]>
| Newsgroups | gmane.linux.utilities.smartmontools |
|---|---|
| Message-ID | <[email protected]> |
On 04/30/15 16:42, [email protected]: > To do so, first you have to figure out the sector number. The short test > gives you the [likely] first number: 673565131. So... > hdparm --read-sector 673565131 <drive; likely /dev/sda> > > Can it read the sector? No? The overwrite it: > hdparm --write-sector 67356131 --yes-i-know-what-i-am-doing <drive; again > likely /dev/sda> Note it may not work for disc with advanced format (e.g. those with physical sector larger than 512B). On some vendor/model/firmware discs block of logical sector covering physical sector needs to be written to invoke relocation. You will got just write error with no relocation otherwise. > Now look at SMART again. Has the pending sectors changed? How about > reallocated? > > Now re-run the short test. Do you get a different LBA number? Yes? Good. > Repeat above for the new error number. > > Keep repeating until all the pending are gone. Short test may not disclose all unreadable sectors. And long test take so much time to be run again and again. In advance, described method require so much user attention. You may prefer unattended mode - just copy complete content of drive to file elsewhere - you can use 'dd conv=sync,noerror ...' for it. Then write the image back. bs= should not be lower than physical sector size. Well, it require other data storage and it take a lot of time. But it can run almost unattended and every sector will be written. Dan ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y