amendment for "Bad block HOWTO" page
Peter van Hoof <[email protected]>
| Newsgroups | gmane.linux.utilities.smartmontools |
|---|---|
| Organization | Royal Observatory of Belgium |
| Message-ID | <[email protected]> |
Hi,
I found that the information given on the "Bad block HOWTO" page for
forcing a bad sector to be replaced doesn't work on a disk with a 4 kiB
physical sector size. At least that is the case on my Seagate Barracuda
7200.14 3 TB drive. This is what I found:
Running a short self-test on this drive revealed a bad sector:
# 1 Short offline Completed: read failure 70% 10174
4130265680
I double-checked this with dd:
# dd if=/dev/sde of=/dev/null bs=512 skip=4130265680 count=1
dd: error reading ‘/dev/sde’: Input/output error
0+0 records in
0+0 records out
0 bytes (0 B) copied, 6.80211 s, 0.0 kB/s
So indeed, this sector is bad and I want to replace it. This is what I
did, following the description on your (otherwise excellent) HOWTO page:
# dd if=/dev/zero of=/dev/sde bs=512 seek=4130265680 count=1
dd: error writing ‘/dev/sde’: Input/output error
1+0 records in
0+0 records out
0 bytes (0 B) copied, 22.6075 s, 0.0 kB/s
The operation failed! Doing another read operation as shown above
confirmed that the bad sector was indeed not replaced. So on a hunch I
decided to try using a 4 kiB buffer size. This means that the value for
the skip parameter needs to be adjusted. There are eight logical
512-byte sectors in one 4096-byte physical sector, so the new skip value
should be 4130265680 / 8 = 516283210. Using this, the new command becomes:
# dd if=/dev/zero of=/dev/sde bs=4096 seek=516283210 count=1
1+0 records in
1+0 records out
4096 bytes (4.1 kB) copied, 0.000359402 s, 11.4 MB/s
And now the operation succeeded! Double-checking confirmed that the bad
sector was indeed replaced now:
# dd if=/dev/sde of=/dev/null bs=512 skip=4130265680 count=1
1+0 records in
1+0 records out
512 bytes (512 B) copied, 0.000310025 s, 1.7 MB/s
So it looks like replacing a bad sector only works if you overwrite the
entire physical sector in one single write operation. Overwriting only
part of the physical sector has no effect.
Peter van Hoof.
------------------------------------------------------------------------------
Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports
Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper
Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer
http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk
_______________________________________________
Smartmontools-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/smartmontools-support