Is My Hard Drive Failing or Not?
I have a 500 GB Seagate hard drive that's about 3-1/2 years old. It only has 16.88 GB of data on it. PCLOS was running very flacky so I popped in a live disk of Vector. When it booted, S.M.A.R.T said that the hard drive was failing so I backed up everything. Since then I haven't written anything to the hard drive. I also blew out a lot of dust from inside the case.
Today I booted a live disk of Salix OS. Ran fsck, badblocks and S.M.A.R.T. and got conflicting results. fsck and badblocks say the hard drive's OK, but S.M.A.R.T. says it's going to fail within hours. umount /dev/sda gives the message that it's unmounted already. But then doing fsck on /dev/sda gives the message that the drive is mounted or busy. /sda1 is the large partition on /dev/sda where I store my data. Can anyone interpret these outputs? Let me know what other information you need. As usual, I'm just lost ... ++++++++++++++ Code:
root[one]# umount /dev/sda Code:
# umount /dev/sda1 Code:
# badblocks -v /dev/sda1 Code:
# smartctl -a /dev/sda1 Thanks bunches! |
Please use code-tags for output of programs in the future, it makes your posts much more readable and preserves the formatting.
To your problem: That you can't mount/unmount /dev/sda is pretty normal, since you state that it is partitioned. You can only mount file-systems, not whole partitioned disks. The smartctl output indicates that the disk is dying and I personally would trust that program. To be sure you can test the disk with the manufacturer's diagnosis tool. |
Where are the instructions for code tags? I honestly looked but couldn't find them.
Thanks for the advice. |
If you use the advanced editor you can just mark the part of your post that you want to put into code-tags and press the #-button above the input field.
You cans also do it manually with writing the tags , for example [code]lspci[/code] will look like this: Code:
lspci |
Thanks, TobiSGD. Fixed it.
|
Yes, the drive is dying, so replace it.
You can also run SMART long tests to check for bad blocks, but if an attribute is already failing, there is no use. |
Quote:
I just bought a Samsung 128 GB SSD and another Seagate 500 GB HDD. I'd like to test them before putting data on them. What are the best ways to do this? |
I always run a long test after I buy it and at periodic intervals (say 1000 power on hours).
Code:
smartctl -t long /dev/sda |
To TobiSGD,
why would not trust the output of smartctl? Call me cynical but I would rather doubt the results of the manufacturers test disk as it would surely give favorable results, no? I've found the smart test on partedmagic to have accuratly predicted a drive's demise several times.It certainly gives more details than some test disks from some manufacturers. Fred. |
Quote:
Quote:
|
I tend to trust the OEM's extended drive tests. They have more insight to their own equipment. I doubt they'd risk an intentional error that made their product look good. Smart is most of but not the entire diags being run. The long test is the best test to use for hard drives.
|
All times are GMT -5. The time now is 09:54 AM. |