Welcome, Guest
Username: Password: Remember me

Search Micromat Forum


TOPIC: Surface Scan

Surface Scan 19 Apr 2011 02:00 #373

I used TT Pro 5 and ran a surface scan. I found bad blocks. I backed up and did a zero erase with Apple Disk Utility. I reinstalled the OS and now I find more bad blocks. They look like the same block numbers that were found before. If I read the docs correctly this should not be happening. I suspect the drive is going bad, but the S.M.A.R.T. status is good. Any suggestions on making sure the bad blocks are not used for data?

iMac 24" 1TB Hitachi drive.
The administrator has disabled public write access.

Re: Surface Scan 19 Apr 2011 03:03 #374

The S.M.A.R.T. status will continue to say that the drive passes until the number of spare blocks (blocks set aside for remapping bad blocks) is dangerously low.

The drive should be backed up and replaced. The magnetic media is going bad. I will post the lengthy discussion of bad blocks from our original forums later.
MicroMat Inc
Makers of TechTool
Last Edit: 29 Sep 2011 08:20 by micromattech3.
The administrator has disabled public write access.

Re: Surface Scan 19 Apr 2011 04:51 #377

If I had not backed up this detailed discussion of bad blocks into my Yojimbo database, I could have found the cached version at Google by searching at Google for:

"unremapped bad blocks" Christmas

The cached version shows this text in its original context, an optimization that failed because of bad blocks.

[Revised Nov. 21, 2014 ]

A bad block is a block that can no longer be read or written to, because the magnetic media has flaked off the drive’s platter. If the block contained data, the data is gone. The drive may still be able to be used if you make a backup of all the files, then reformat the drive with the option to zero all data chosen, using Disk Utility in Mac OS X 10.2.3 or higher. (The Drive Setup utility in Mac OS 9 can also perform this type of reformat.) All volumes on the drive and their files are lost in this reformat.

If you need to make a backup of a bootable volume on the drive, be sure to use a utility that can make bootable backups. An installation of Mac OS X is more than a collection of files, and cannot be backed up by the Finder.

I encourage you to try to restore your backed-up files to another drive before you reformat the drive containing the bad blocks and your original data. It is discouraging to make a backup and later receive an error such as “Restore failure: could not open connection to helper tool” when you attempt to restore the files. It is significantly more discouraging if you have already reformatted the drive, making the creation of a new backup impossible.

Most drives contain some bad blocks. When the disk driver encounters a block to which it cannot write, it is supposed to remap the bad block to one of a set of blocks set aside for this purpose, called spare blocks (which we hope are good themselves). The bad blocks that are found and remapped to spares at the factory are called factory defects, and the bad blocks that develop later are called grown defects.

Spare blocks initially contain no data. They receive data when the disk driver attempts to write to the location of the bad block. The data is written to the corresponding spare block.

The reformat with the option to zero all data chosen allows the disk driver a chance to remap permanently any blocks it regards as bad to spare blocks. Every block on the disk is checked and a binary zero is written to each spot that can hold data.

TechTool Pro can detect bad blocks that have not yet been remapped to spares, but it cannot order the disk driver to remap them. Most bad blocks encountered by the disk driver get remapped to spares immediately (“on-the-fly”) in the course of normal use of the drive, but in some cases, the bad blocks do not get remapped during normal use. Today’s ATA/IDE drives are significantly better at remapping bad blocks on-the-fly than old SCSI drives were.

If the reformat with the option to zero all data fails, it is almost always because so many bad blocks have been remapped (either in the past or in the course of the present reformat) that no more can be, and the drive should be replaced. If the reformat succeeds, restore the system software and your applications and files from the backup.

If you are using TechTool Pro 4, be sure to open the Safety window and turn on the Protection feature and the S.M.A.R.T. test. The S.M.A.R.T. test monitors whether the rate at which bad blocks are being remapped exceeds what the manufacturer of the mechanism regards as normal, and whether the number of remaining spares is considered sufficient.

TechTool Pro will not report already remapped bad blocks as bad if you run the Surface Scan, because every request to the read/write head to visit the bad block is met with a trip to the corresponding spare (good) block. If you reformat a drive using Disk Utility with the option to zero all data chosen, and then run the Surface Scan in TechTool Pro and find bad blocks, there are four possibilities:

1. More magnetic media has flaked off the platter. There is no way to know how much more is going to flake off or when, or whether such media will contain data. I do not see how one can trust data to a drive that has shed magnetic media since the reformat.

2. The spare blocks themselves are flaking off. The spare blocks are set aside and cannot be deliberately written to by the user, but there is nothing physically special about their media. In some cases, this condition will rapidly exhaust the supply of spares, and the drive will report a S.M.A.R.T. failure, either because there are too few remaining spares, or the rate at which bad blocks are being remapped to spares is too high.

3. The firmware of the ATA/IDE drive, which contains the table of bad blocks and their corresponding spares, has become corrupted.

4. The computer has been used when it was so hot that the processor was malfunctioning in such a way that the drive could not read perfectly good data from the platter. The failed attempt to read the data from the platter is reported by the Surface Scan as an "error", which is an unremapped bad block. The technical specifications for your Macintosh give you the range of ambient air temperatures in which the computer can function properly. The usual maximum is 95 degrees Fahrenheit, or 35 degrees celsius.

Any block on a hard drive can go bad at any time, because magnetic media can flake off the platter at any time.

If you get a new drive and run the Surface Scan, or you have just run the Surface Scan on a drive that has not been previously examined for bad blocks, and one or two bad blocks are found, then it is likely that a small amount of the magnetic media has flaked off since the drive left the factory. This is not uncommon, and is cause for neither panic nor complacency. (Many new drives have one or two bad blocks that have developed since the drive left the factory, but most users are unaware of them, because they do not have the Surface Scan.) Back up the drive, reformat it with Disk Utility with the option to zero all data chosen, and put the drive into use. Back up its files frequently to another drive (in other words, continue your usual practice of frequent backups), and run the Surface Scan often to see if more bad blocks have developed. The Surface Scan is not a spectator sport; the test can be run overnight.

If the Surface Scan ever finds more than one or two bad blocks, the drive should be replaced.

My new MacBook Pro Core 2 Duo has a 200 GB hard drive. After Apple’s Migration Assistant did a flawless job of moving my user accounts, settings, files, and applications to the MacBook Pro, I ran the Surface Scan and found that one of the 390 million blocks was bad. I backed up the drive using Disk Utility, reformatted it with the option to zero all data chosen, restored the files, and will be running the Surface Scan overnight frequently to see if any new defects are found.

Update: My MacBook Pro Core 2 Duo passed the Surface Scan on Christmas Eve, 2007. On January 7, 2008, one bad block was found. It was in the disk directory. There was no warning from S.M.A.R.T., but because the Diagnostic Scan in TechTool Pro kept causing the computer to slow to a crawl for about 90 seconds, I investigated the All Messages log in Leopard and found that there was an i/o (input/output) error during the Diagnostic Scan:

1/7/08 1:51:46 PM [0x0-0x534534].com.techtoolpro4.daemon[70842] # ReadRawBlock attempt to read 16777216 bytes at offset 278798336.000000 failed with Input/output error
1/7/08 1:51:46 PM [0x0-0x534534].com.techtoolpro4.daemon[70842] # ReadRawBlock CANNOT READ: blockNumber 544528.000000, numBlocks: 32768, errno: 5, Input/output error

I deduced that the bad block was in the disk directory when TechTool Pro and Disk Utility refused to rebuild or repair the directory. That was because the bad block prevented it from being read. I am still trying to find online the report from a customer who wrote that he installed TechTool Pro and his computer kept crawling, and he removed it and it was back to normal. He thinks he solved a problem, but he may have simply ignored an impending disaster.

It appears that the bad block on my disk directory recorded the locations of read-only system files that were backed up by Time Machine before the bad block arose. I lost no data at all when I restored from the Time Machine backup on to the new drive.
MicroMat Inc
Makers of TechTool
Last Edit: 21 Nov 2014 22:53 by micromattech3.
The administrator has disabled public write access.
The following user(s) said Thank You: Ian Goddard, JHFowler

Re: Surface Scan 18 Nov 2013 20:29 #4938

Thank you for all of your help and, especially your patience! I have learned a lot.

I wish I could say the same for LaCie. However, in fairness the drive is 3 years old. I am beginning to see alternatives for backup such as the cloud which can solve much frustration.


Ian Goddard
The administrator has disabled public write access.

Re: Surface Scan 18 Nov 2013 20:59 #4939

You are welcome, and thanks for your kind comment.

Given the fact that all man-made objects have failure modes, I think the best approach is to consider supplementing hard drive backups with a cloud-based backup. You need at least one backup that is off-site, in case the location of your computer is destroyed. The onsite backups should not all rely on the same method. I have both a Time Machine backup, which is very useful until its complex disk directory gets damaged, and a clone made with Super Duper. Each of these programs has briefly failed for me in the past, but fortunately, they did not both fail at the same time.
MicroMat Inc
Makers of TechTool
The administrator has disabled public write access.
Time to create page: 0.371 seconds