Hard Disk Monitoring

Hi HDS,

I recently can across an issue which HDS didn't resolve in a way I had hoped it would - I'm posting this more as an experience than a question.

I have a 8-way RAID5 running on a Highpoint RocketRaid 2680 - HDS supports and shows all the details on all the drives attached to this perfectly.

Recently I had a drive failure - instantaneous - nothing HDS or anything could have helped. So I put in a new drive, and the rebuild began.

Unfortunately, another drive began to fail with bad sectors, preventing the rebuild from completing. So with the RAID array offline, I got HDS to scan the individual drive, and yes 2 sectors were reported. However, no matter how many times I got it to scan that section of the disk with a refresh data surface scan, it would not remap the bad sectors. This was disappointing, as I thought HDS would do this. HDS will remap bad sectors perfectly on this controller using the Initialise Surface scan test - but naturally I want to keep the data, to be able to complete the rebuild.

I have read in one post elsewhere on this board (back in 2010) that someone else found that the refresh data surface scan did not invoke a bad sector cleanup. viewtopic.php?f=18&t=789&p=1106&hilit=refresh#p1107

In the end I had to revert to booting a DOS USB with an ancient (but utterly indispensable) surface scanner Vivard from copyrsoft.com which remapped the bad sectors immediately without wiping the rest of the HDD. Upon retesting with HDS running the refresh surface over that same part of the HDD, all sectors were now perfect.

This allowed me to rebuild my RAID array without further issues.

I don't know whether the failure of HDS to remap the bad sectors during the refresh data scan was because of the HDD, the Controller, or something else - but it was somewhat alarming to only find out this limitation at the exact moment I needed it to function correctly, and remap the bad sectors without wiping the rest of my drive.

Many thanks
Steve

Hi,

I can confirm that yes, as you wrote the "Refresh Data Area" test may not force re-allocation of bad sectors in all cases.

This is not really a limitation of Hard Disk Sentinel, but there are certain factors which can cause this:
exactly as you wrote, the hard disk itself, the hard disk controller (or more precisely, its driver) and the OS.
This is why things may work under different OS and/or if the drive is connected to a different controller.

This is why the destructive test (Reinitiliase Disk Surface) is the best method to force the reallocation of problematic sectors, as described in the description of the test, and as your tests confirm that.

Of course I completely understand the situation and we're constantly looking for ways to improve the efficiency, so I can confirm the "Refresh Data Area" test is also improving to repair problematic sectors (re-allocate if required) in different configurations, to perform the re-allocation without affecting data stored on good sectors.

If you have the opportunity, please use Report -> Send test report to developer option as it always help to investigate the possibilities and improve the software. Thanks!