Slow External Hard Drive w/ Bad Blocks
I bought a LaCie d2 Quadra Hard Disk 1TB drive about four months ago and was home from school for Christmas break and noticed it was extremely slow. I noticed it because I was doing large file transfers between my computer, my sisters computer, and my two external drives; the LaCie and a four year old Western Digital.
The LaCie had about a 55 megabits/sec transfer rate over Firewire 800 where as the WD had a transfer rate of 180 megabits/sec over USB 2.0 (only port on it).
So I tried different cables and different ports seeing if it was something simple and nothing made a difference. I also tried the USB port on this drive and got about the same 55Mb/s.
So I ran a full computer test using TechTool Pro 5.0.6 and found that the LaCie has 6 bad blocks and I am wondering if this is what could cause the drive to be so slow.
Both external drives are formatted the same: GUID Partition Table in Mac OS Extended Journaled.
Also, the LaCie is formatted into three partitions: one for files, one is bootable, and the other for time machine and none of the partitions are even close to full. I also ran a free space consolidation or w/e it is called to make sure it had plenty of free blocks to write to and that did not make any difference.
Any help or suggestion to speed up the drive would be greatly appreciated, or a better explanation of bad blocks and if that could cause the drive to be as slow as it is.
I am willing to reformat to see if that helps if someone thinks it will, but am not very excited to do so.
Thanks,
Robert
The LaCie had about a 55 megabits/sec transfer rate over Firewire 800 where as the WD had a transfer rate of 180 megabits/sec over USB 2.0 (only port on it).
So I tried different cables and different ports seeing if it was something simple and nothing made a difference. I also tried the USB port on this drive and got about the same 55Mb/s.
So I ran a full computer test using TechTool Pro 5.0.6 and found that the LaCie has 6 bad blocks and I am wondering if this is what could cause the drive to be so slow.
Both external drives are formatted the same: GUID Partition Table in Mac OS Extended Journaled.
Also, the LaCie is formatted into three partitions: one for files, one is bootable, and the other for time machine and none of the partitions are even close to full. I also ran a free space consolidation or w/e it is called to make sure it had plenty of free blocks to write to and that did not make any difference.
Any help or suggestion to speed up the drive would be greatly appreciated, or a better explanation of bad blocks and if that could cause the drive to be as slow as it is.
I am willing to reformat to see if that helps if someone thinks it will, but am not very excited to do so.
Thanks,
Robert
Comments
Unfortunately it is really difficult to tell the first case from the second. So you pretty much have to assume that it is the second. At the very least it is time to get the data backed up if you have not already done so. And I mean quickly.
Once you have backed up the data (so the drive can die if it likes), and if you want to play with fire (as opposed to just assuming the drive is bad and replacing it... a safer assumption), then if you want to try and rescue the drive (but not the data you have already backed up) you should reformat the entire drive. Repartition it to one partition and do a format writing zeroes to the whole drive. This will take a long time (even on a good drive, bad drives will take forever, and seem to pause at a lot of places), but this gives the drive the best chance of mapping out bad blocks.
On the whole my thought it that it is not worth the time. Replace the drive and get on with things.
...you should reformat the entire drive. Repartition it to one partition and do a format writing zeroes to the whole drive. This will take a long time (even on a good drive, bad drives will take forever, and seem to pause at a lot of places), but this gives the drive the best chance of mapping out bad blocks.
Get your data onto another drive first and then Disk Utility Erase with Zero option on the problem drive, it will take a hour for 350GB or longer for larger drives. This will map off the bad sectors, even the ones not found yet.
Don't count on this drive until it has proven itself reliable, use it as a mirror or a second backup.
Sometimes rough handling causes drives problems, but just limited to bad sectors, but it also could be a mechanical or electrical problem and that means it's worthless.
I as a matter of routine, always Zero every new drive I get before placing data on it, even new Mac's. I also buy the very best drives from Hitachi.
My reliability factor has tremendously increased and only my Apple drives have failed, but I plan on that by cloning my boot drive regularly. Make Apple fix their poor drive maker choices.
You can learn how to clone your drive in the sticky thread at the top of this forum.
Also, just thought I would let you know that it took about 26 hours to zero out the drive
No hard drive should have bad sectors. If it does, then you have a bad hard drive and should replace it asap
Where did you get that idea? At a low level almost all drives are going to have bad sectors in their productive lifetimes, that is why drive manufacturers dedicate a portion of the drive to be spare sectors. When a production sector goes bad (as determined by the drive's firmware) an unused sector from the reserve space gets mapped into its place. If things go well the computer should never even know that this has happend.
Sometimes a sector goes out and takes something important with it (like filesystem metadata), and that is one of the causes of disk corruption. If the data it takes with it is file data, then that is file corruption.
This is all very normal, and taking place below the layers you are used to seeing. This is one of the reasons that people spend a lot of money on RAID systems that read-back the data from multiple drives at the same time (and then compare the results). The venerable points on this system are exactly why ZFS is such a big thing with people who have really large amounts of data (and thus are statistically guaranteed to be caught by these problems). It is the reason that Sun is throwing the money into ZFS, and Oracle into BTRFS. Both have checksumming to prevent exactly these problems (and others) from becoming data problems.