First internal SSD failure
Rsadeika
Posts: 3,837
My desktop system, Gigabyte mini-itx board, Intel iCore3 3.3 GHz processor, 16 GB RAM, Samsung SSD 850 EVO, and Ubuntu 15.10. The symptoms are very interesting.
Basically, from a cold start, it gets through the BIOS stuff, like loading an OS and gets to the screen where it shows the Ubuntu in the center with the flashing big dots, and that is where it gets stuck. It does not go any further than that. I was expecting it to stop when it tried to load the OS, but it gets passed that.
To try too figure out what was going on, I ended up starting a Mint live disk. After about 5 minutes of CD drive thrashing, it finally started up. Next, checked to see if the SS drive was there, Mint showed the drive. When I tried to read the drive, again, it took awhile, but it finally showed the contents, with a note that some of the files and folders could not be read.
Now I am thinking that the boot sector was still somewhat viable, but the kernel or some parts of the files that are needed are corrupted, or the actual sectors on the drive, that hold that stuff, are dead.
I also was digging around for a standalone program that could check the drive, and maybe fix it?, but no such thing is available for the Linux stuff. And I was led to believe that Linux had every tool imaginable at your disposal. I guess, at this point, I could try to install Ubuntu again, and see how long it takes to get to next failure.
I am not sure if I want to turn this into a test box, checking how long between failures. The Samsung SSD package did come with a CD, but it is not a standalone, and I believe it needs a Windows OS to work. If I do get another SSD, I have to figure out which company makes the most reliable ones, or are they all more or less the same, from a reliability standpoint? Not sure what to try next, before I yank and toss.
Ray
Basically, from a cold start, it gets through the BIOS stuff, like loading an OS and gets to the screen where it shows the Ubuntu in the center with the flashing big dots, and that is where it gets stuck. It does not go any further than that. I was expecting it to stop when it tried to load the OS, but it gets passed that.
To try too figure out what was going on, I ended up starting a Mint live disk. After about 5 minutes of CD drive thrashing, it finally started up. Next, checked to see if the SS drive was there, Mint showed the drive. When I tried to read the drive, again, it took awhile, but it finally showed the contents, with a note that some of the files and folders could not be read.
Now I am thinking that the boot sector was still somewhat viable, but the kernel or some parts of the files that are needed are corrupted, or the actual sectors on the drive, that hold that stuff, are dead.
I also was digging around for a standalone program that could check the drive, and maybe fix it?, but no such thing is available for the Linux stuff. And I was led to believe that Linux had every tool imaginable at your disposal. I guess, at this point, I could try to install Ubuntu again, and see how long it takes to get to next failure.
I am not sure if I want to turn this into a test box, checking how long between failures. The Samsung SSD package did come with a CD, but it is not a standalone, and I believe it needs a Windows OS to work. If I do get another SSD, I have to figure out which company makes the most reliable ones, or are they all more or less the same, from a reliability standpoint? Not sure what to try next, before I yank and toss.
Ray
Comments
I won't touch on the "and maybe fix it" because that depends on the problem. Don't know what to search for until you know what the problem is. However, for discovering the problem, a quick Google showed this page: http://askubuntu.com/questions/325283/how-do-i-check-the-health-of-a-ssd
Have you tried any of those tools?
PS - Sorry, should have said ssd diagnostic tool. Old habits die hard.
That software is great, I'm surprised you have to dig around for Linux support for it.
I'd speculate that you've managed to somehow corrupt the filesystem is all. A fresh install of Ubuntu may be the easiest fix.
Also, before doing anything else, if you have a large unused external drive with enough free space, boot off of a live CD and use dd or ddrescue to copy the whole disk into a file or partition on the external drive.
When I had the live disk running, the Disk Usage Analyzer showed that the SSD was ~245Gb available, and ~84GB used. After doing the re-install, it is now ~235GB available, and ~15GB used. So, it looks like about 80GB of files was lost.
I did a re-install of Wine and tried the Samsung Magician disk, it would not run with Wine, now I will have to search some more for an SSD analyzer program. I guess this will be a test box.
I also have my USB StarTech box which has 4 - 2TB drives attached, now I will see what is left on the re-install that can be saved to that. This is one aspect of Linux that I do not like, your home folder is on the SSD as a default install. At least with Windows, you can just have the OS on the SSD, and everything else goes to a different drive.
Ray
The Samsung Magician proggie gets hung up trying to phone home when on Wine. Err, that was printer diag tool I tried. I don't think I've even tried Samsung's proggie.
Also, I did the Tests, and it came up with no errors(Passed). The other thing that I noticed is that it shows 238.47 GB capacity, while the Disk Usage Analyzer shows 235.1 GB capacity, whats a few GB difference, after all it is a desktop Linux.
But the biggest problem is, what the heck happened? I shut the machine down last night, and this morning on startup I get the equivalent of the "blue screen of death". Yesterday when I was using the system, their was no indication that something was wrong. Where is the self healing Linux programs when you need them. If this happens again...
Ray
PS: I'm using an 840EVO for my main drive.
PPS: Intriguing that you get about 5.5GB more free space on the 850 compared to the 840. Samsung have obviously adjusted the amount of reserved space given the total amount of physical flash memory is identical at exactly 256GB. Maybe that's a sign of how fast they expect the flash to wear.
Electrodude's instruction gives the regular tool for volume repairs. It's too late for that now as you've already reinstalled but if it happens again you could investigate. You might get a good indicator of what is causing the problem as the repair tool will say what files are problematic.
You should read about manual partitioning and /etc/fstab. Putting the home partition on a separate drive is one of the easiest things to do on Linux when you install it. Almost all Linux distros recommend or at least suggest that you do this when you install them and usually make it very easy to do so. I would also personally recommend using anything but Ubuntu - it is the Windows of Linux and is very closed and often hard to fix.
Also, only reinstall your OS as a last resort, and don't trust it claiming that it will preserve all of your data (as you just learned). I've fixed every Linux problem I've ever faced (that didn't have to do with a bad installation to begin with) without doing a reinstallation.
I use Kubuntu but I don't have any complaints with Ubuntu's behaviour in general. Distro wars a brewing is it? ... :P
Hello!
Ray, where did you buy the thing? As with everything else you might want to check to see about your device warranty from the vendor, Samsung, and the one from where you bought it from. For example, 95 percent of what I buy branded by Parallax comes from Micro Center stores, any of three in New York State.
Remember the technology is still growing up, there's a chance you might have gotten a duff ( definitely trying to be a failure) one.
Ray has already confirmed the drive is in good health.
Remember I suggested that he might have ended up with a duff one. The keyword was "might". Obviously the thing worked so the big problem is now working out which part of of it did do that.
Relabeling components and goods are so very common, I have seen plenty of electronic components with the right label but the tests reveal something completely different. I would accept that you know for a fact that the label says 850 series but I would not discount what the program is telling you. It is after all not looking at the label but at the actual chip IDs etc.
In regards to Linux installs I have mentioned before about using USB Flash drives instead of CDs as CDs are so so slow, It takes about the same amount of time to "burn" a Flash drive as it does to burn a CD, but the Flash drive slips onto my keyring and fits any computer, even ones without CD drives. Also when it comes to installing on a computer I prefer to live boot and run gparted to setup my 20GB Linux system partition, another 20GB for playing with other distros, about 10GB or so for a swap. and the rest can be home partition. THEN I install Linux selecting "something else" and click the options.
I think we can trust the internal error tracking of the drive itself to know when it has had any issues.
Probably so. No actual model string in the drive's responses. The enquiring program just makes something up for human convenience.
Even SD "controller" chips respond with all kinds of ID and manufacturing information, I have used this to glean a lot of relevant information from various cards. Likewise SSD memory has SSD controller chips too and although I haven't looked into the details of these I would not expect them to be less so than SD controllers.
If I had performed the same tests as Ray I know that I would then be questioning the authenticity of the product itself, or perhaps the supplier. Especially since there seems to be problems with the drive. A genuine drive would not report back as another model.
I'll have a look at how to display the reported ID numbers when I get home ...
I still like Ubuntu. I think I still like Linux. I still like Ubuntu...
Their is a question that has been asked, and the answer is, this particular unit, I put together myself. So, I am very sure that the parts I used were genuine, like for instance, the Samsung SSD. I purchased those from NewEgg, which came in the original packaging.
Thanks for all the responses, I am still determined to figure out what exactly happened to jam up my system. Although I do remember an incidence with, I think it was Mint 16, but, on an off the shelf tower that I had. I remember I installed the OS, which was working as expected, and then one time, on a cold start, the desktop window shrunk down, and the system kept rebooting itself. I never figured that one out. After that incident I installed other OS's, and those worked without any problems.
Now I have to check and see if PropellerIDE and SimpleIDE are still there and functional, although spincvt got carried over and is functional, after I re-installed tklib, that is.
Ray
Try running the memory test from the boot menu. You're not overclocking at all are you?
Oh, I had to download PropellerIDE, and I got the "Internet connection" window pop up, again. Not sure why this is occurring, my other machine does not have the "Internet connection" problem, so is there something wrong with the Ethernet on my motherboard? When I installed Ubuntu, I am making an assumption the installation process was correct for this.
I am starting to get a bad vibe with all of this, not sure if it is software or hardware issues.
Ray