Linux always telling me my drive is failing
I wonder if anyone can tell me why Linux (any version) always tells me my drive is failing? I have bought many new drives over the years based on Linux recommendations.
Now it is telling me a drive is failing yet again. The drive just ran out of warranty the 23rd (seems to always happen that way). Anyways, smartmontools is telling me some things are pre-fail and some are old age. My power on time doesn't even come up to a year. Are drives getting this horrible really? Seagate (according to Linux) I had 3 drives fail and replaced them after only a little over a year use. Now this WD is failing. Horrible. Here is smart info: Code:
smartctl 5.41 2011-06-09 r3365 [x86_64-linux-3.2.0-36-generic] (local build) |
Just plugged in an SSD I had in a laptop. Bought about 6 months ago (Intel). Guess what? It's failing as well.
I guess this can be chalked up as hogwash. I know for a fact the SSD is good (I use the Intel toolbox under Windows to check). So, thanks for looking at this thread. I will be from here on out ignoring the results or the recommendations of Linux on failing drives. |
Bah. To me, drives are cheap and data is priceless. If SMART's telling me that drives are failing, I'm not going to be the one to go on assuming that everything's okay until one day I fin
click click .. click click .. click click .. click click .. youre scrood .. click click .. click click .. ;) |
Quote:
Data (drives) are cheapish. But a failure a year? For home use? Makes no sense. These drives are getting to a point where the quality has diminished so much that they require a yearly replacement? 80 bucks a pop. My machine is getting to be around 4 years old. I paid about 360 bucks for my home build (not including drives). This is costing me 310 bucks in drives and if I act now it will cost me 390? In drives? Really? This is crazy. I had an old 25MHz machine up till about a year ago with the same 12 megabyte drive it started with. I have backups, not scared of failure really. Just a bunch of BS in my book. I have a 3 year old laptop on the same stock drive. Funny thing is after a year of use Linux told me the drive was failing. I backed it up and have never had an issue. Silly scare tactics to make one run out and buy a drive? Linux not reporting info right (in the ssd case I know it is wrong anyway)? Who knows... |
What makes you think that smart says you drive failed
Quote:
|
I agree with whizje--what in the results indicate the drive failed/is failing? The OLD_AGE and PRE_FAIL are just categories the metrics fall into, if you look in the WHEN_FAILED column none of the tests have ever failed.
|
Thank you guys.
Now that has taught me something. I have always looked at that data and replaced the drives. So this has always been a matter of me looking at the information wrong. So now I know I have probably over spent over the years lol. Thanks so much for clearing this up. |
Why would you replace the drive before it fails anyway? I take smart warnings like that as just what they are, a warning sign..."hey your drive is about to fail, so make sure you're current on backups and have a replacement ready to go". Then when it does actually fail, a year or two later (if ever), everything is ready to swap in.
|
The first time I ran smartctl, on a brand new drive that I just partitioned, I saw "pre-fail" and "old-age" I thought really that the hard drive was failing, as I had problems with another drive (WD20EARX) a few days previously. However, I remarked that there were no fail event really reported and I then understood better the report.
Anyway, it is not usually possible to predict drive failures. - A friend recently lost 2 drives (hmmm... WD green too) in a NAS RAID-5 array. Normally, the controller checks SMART and warns about the drive status. In this case, there was no warning. - Another recent case, in a forum server: failure of a drive in a RAID-1 array, failure of the second (good) drive during RAID rebuild! In both cases restoring the backup was the solution. Backup, backup again, backup often. |
Well, just an update.
Drive failed. Had some problems the other day (read errors and such). Got a new case and transferred everything over to the new case (old one was a dust magnet). Booted the machine up and it started without issue. Couple updates installed, rebooted. Sat and stared at the screen for 10 minutes waiting for something. Dropped to console to watch the boot process, read write errors everywhere. Let it run for over an hour and nothing. Drive is gone. So now running SSD with a 320GB drive I had around mounted under /home/me/media for my bigger files and got cache and swap setting there. Much faster but lets see how long this lasts. I guess my paranoia had something to it. Dropped my old drive into me new esata port and checked smart, passed. So smart must not be so smart. Oh well... |
SMART data can be interpreted - different drives can report certain values improperly.
But generally you need to look out for these values - these tend to be reported by most drives correctly: Code:
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE If you can access the drive, you can try zeroing the whole drive a few times in a row ("dd if=/dev/zero of=/dev/sdX", where X is a,b,c,etc corresponding to the drive in question). |
Quote:
Here is the result of the first run: Code:
dd: writing to `/dev/sdc': Input/output error And again: Code:
dd: writing to `/dev/sdc': Input/output error |
The info you posted for the first drive suggests that the drive is fine.
If you just bought a new drive and it is failing somehow, I would not suspect the drive. |
A Current_Pending_Sector count of 124 is hardly "fine". Something bad has happened to that drive, either during the 4 hours since that last successful short offline test, or something that the short test does not detect.
|
Quote:
If you really are having hard drives fail on you this often, I would take a closer look at your setup. I used to lose a drive a year on my personal computer until I started putting a case fan on the drive (just a case fan on the front of the case in front of the drive to give it some air flow). I haven't lost a single drive on my personal computers (which currently account for 3 comps and 9 hard drives out of the list above) since I started doing that about 9 years ago. |
All times are GMT -5. The time now is 08:20 PM. |