Please report all spam threads, posts and suspicious members. We receive spam notifications and will take immediate action!
Page 3 of 3 FirstFirst 123
Results 21 to 26 of 26

Thread: Real problem, corruptions and disks inaccessable - eek! GA-EX58-UD5 board




  1. #21
    Join Date
    Apr 2009
    Location
    J'Habite En Angleterre
    Posts
    2,431

    Default Re: Real problem, corruptions and disks inaccessable - eek! GA-EX58-UD5 board

    It's not "Guessing" it's standard trouble shooting. SMART data isn't some sort of voodoo or a magic bullet. It's not even 100% foolproof. What is the OP to do if all SMART values report as normal?

    It's also perfectly possible for one program not to be able to read SMART data and yet another one may be able to without issue. Passing through of SMART data by the controller is not an issue. We know what controller he's using and we know that the ICH10R passes SMART data through just fine.

    BTW, increasing SB voltage has nothing to do with voltage signals sent to/from the drive. The SB is an IC just like the NB and CPU and when stretched with a drive on each, a voltage increase can solve a number of issues. I don't care how likely you think it is, this is from personal experience. If a SSD is transfering at 140MB/s sequential read at stock SB voltage and with an increase of 0.1V the speed jumps up to 210MB/s, then that's insufficient SB voltage causing a problem. And yes, decreasing voltage back to stock caused performance to drop again. I see no reason why this can not apply to data integrity too. It's not just at the drive level or at the data cable where data can get messed up.

    Sure, looking at the SMART data may be a good idea but dismissing basic troubleshooting steps as "guessing" is extremely unwise.
    Coolermaster CM 690 II advance Case
    Corsair HX750 (CWT, 91%(80+ Gold rated @230V) single 62A 12V rail
    P55A-UD4 v2.0 @ F14
    Core i5 760 @ 20 x 201, 4.02GHz
    TRUE Black with a single Noctua NF-P12 pumping out 55 CFM @ 19db .
    2 x 2GB Mushkin Ridgeback (996902), @ 7-10-8-27, 2010-DDR, 1.66v
    2 x Gigabyte GTX 460 1024MB in SLI (Pre OC'd to 715MHz core and 1800MHz VRAM) @ 850 Core / 4100 Mem.
    Intel X25-M Boot Drive (OS and Programs) 200MB/s Read & 90MB/s Write
    Corsair X32 200MB/s Read & 100MB/s Write
    WD Caviar Blue 640GB C (Steam, Games, Storage, Temp Files & Folders, etc)
    Samsung F3 500GB Backup/Images
    Noctua 1300RPM 19dB case fan (rear extraction)
    3 x 140 MM Coolermaster LED fans (one front intake, one top extraction, one side intake)
    Dell Ultra Sharp 2209WAf E-IPS @ 1680x1050

  2. #22
    Join Date
    Jan 2010
    Location
    Someplace
    Posts
    17

    Default Re: Real problem, corruptions and disks inaccessable - eek! GA-EX58-UD5 board

    Those read/write failures would be translated as Read error rate, CRC and ECC errors. Data doesn't doesn't goto the bit bucket just because the voltage levels drop on the serial bus. This is why I opted for data over making changes without substantial evidence.

    If you want to conjecture over the next stage of procedures before actually acquiring data we can discuss this. We have a system which was functioning and then ceased to do so. Why? Well it's silly to guess when we have perfectly useful tools, but there are so many components involved. Disk, controller, cable and power. I would opt to blame the controller or if we suspected power levels it's perfectly normal to reduce the complexity of the system and retest. (ie, remove a disk or test the disk outside of the system)

    On an aside, likely in your described scenario you had a higher then normal read error rate. If you are over taxing the voltage regulators I could see such an issue arising. Increasing the voltage would accommodate the failures and likely decrease the read error rate. All of this is perfectly interrupted through smart data.

    A quick tip... when dealing with multiple drives in a raid set I also find it useful to take the media deviation when isolating failure and no useful data can be found.

    I don't really feeling like discussing SMART implementation or the use of the SAT opcode to explain why it works and doesn't. I'll end with the application works perfectly find on my system and is based on the smartmon toolset. An actively maintained project which supports a vast array of configurations.

    I find increasing voltage levels randomly extremely unwise.

  3. #23
    Join Date
    Apr 2009
    Location
    J'Habite En Angleterre
    Posts
    2,431

    Default Re: Real problem, corruptions and disks inaccessable - eek! GA-EX58-UD5 board

    I'm not interested in entering into a discussion with you. I've given my advice to the OP, it's now up to them what they do. I'm glad you don't want to discuss SMART implementation too because I'm really not interested. There have been no problems that I haven't been able to solve as of yet. Apart form isolated instanes where it was obvious from SMART data that the drive was struggling to spin up, every other HDD issue I've encountered has been solved via checking cables, testing PSU's, tuning SB voltage while keeping in the approved range (not "randomly increasing it" etc.

    Slightly increasing SB voltage is not harmful at all, and frankly the point of "overtaxing voltage regulators" is just rubbish.

    A SB requiring more power can be perfectly normal, depending on how many disks are hooked up to it and how much data is passing through it. The Intel SB can operate at a voltage of between 1.1V and 1.35V according to Intel specs, and the default for this board is 1.1V, possibly not enough, who knows.

    When all is said and done, SMART will only possibly tell you what symptoms the disk is experiencing, it still takes practical solutions to correct the problem.

    As for data not being corrupted by insufficient SB voltage, grab yourself a resistor of appropriate size to reduce your SB voltage to lets say .08V and see what that does for your data integrity.

    Talk about using a sledgehammer to crack a wallnut...
    Coolermaster CM 690 II advance Case
    Corsair HX750 (CWT, 91%(80+ Gold rated @230V) single 62A 12V rail
    P55A-UD4 v2.0 @ F14
    Core i5 760 @ 20 x 201, 4.02GHz
    TRUE Black with a single Noctua NF-P12 pumping out 55 CFM @ 19db .
    2 x 2GB Mushkin Ridgeback (996902), @ 7-10-8-27, 2010-DDR, 1.66v
    2 x Gigabyte GTX 460 1024MB in SLI (Pre OC'd to 715MHz core and 1800MHz VRAM) @ 850 Core / 4100 Mem.
    Intel X25-M Boot Drive (OS and Programs) 200MB/s Read & 90MB/s Write
    Corsair X32 200MB/s Read & 100MB/s Write
    WD Caviar Blue 640GB C (Steam, Games, Storage, Temp Files & Folders, etc)
    Samsung F3 500GB Backup/Images
    Noctua 1300RPM 19dB case fan (rear extraction)
    3 x 140 MM Coolermaster LED fans (one front intake, one top extraction, one side intake)
    Dell Ultra Sharp 2209WAf E-IPS @ 1680x1050

  4. #24
    Join Date
    Jan 2010
    Location
    Someplace
    Posts
    17

    Default Re: Real problem, corruptions and disks inaccessable - eek! GA-EX58-UD5 board

    I never said data would not be corrupted by reduced voltage on the serial line. I said it would be perfectly and clearly identifiable.

    Obviously if applying haphazard settings works for you then please continue to do so. However, I have to consistently provide verifiable steps and data to substantiate any failure claims I make.

  5. #25
    Join Date
    Apr 2009
    Location
    J'Habite En Angleterre
    Posts
    2,431

    Default Re: Real problem, corruptions and disks inaccessable - eek! GA-EX58-UD5 board

    You seem not able to understand that ICH voltage and I|CH I/O voltage are two sperate things.... Increasing ICH voltage does NOT increase anything to do with SATA cable voiltage, it simply increases power to the SB IC itself, nothing external gets an increase in voltage.

    The OP needs to use every tool available to them to solve their problem. Where in my posts did I ever say to completely ignore smart data?

    All you have done is say SMART, SMART, SMART. In none of your posts have you either offered possible explanations or signs in the SMART data that the person should be looking for.

    This forum is nothing like your job. People come here asking questions and very often get bored half way through for one reason or another and no longer follow the thread. T|hey need to be able to know the possible remedies for any given problem. If someone's experiencing data corruption it's for one of three reasons in the main:

    1) Drive is fu**ed
    2) Hardware settings or physical connections to other HW such as cables are fu**ed/need changing
    3)Software problems.

    A quick glance at the drives SMART data may be useful, sure. But they don't have to logg everything in triplicate, they need to find the problem and fix it. And I repat again, SMART data is NO|T infalable. It is a function provided and controlled by the drive's firmware. Poorly programmed FW CAN lead to unreliable, sometimes even garbage data being reported. Apparently both my Indilinx SSD's for example are flagged as being "Critical and their "Average Erase Count is well above 10,000. This was from brand new. Several WD2000JS drives have had obvious dive motor problems, evident from being able to hear the motors struggle to spin up. Funnily enough, your much coverted "be all and end all" SMART data showed not a blip, all values were in normal ranges and CrystalDiskInfo happily reported the drive condition as "Good". Clearly it was not.

    Given that any conclusion that may be possibly gleened by using SMART (not definitely) will require steps to be taken similar to if not identical to those I have advised then why is this "Haphazard"? SMART data has been shown in various studies including one carried out by Google on 100,000 drives to only be 64% effective meaning that in 36% of cases the SMART data was useless. In all honesty, talking about in the consumer market, those odds are deceptively close to a 50/50 "punt" that you're going to be seeing any useful data pertaining to the current issue. Changing a cable, checking the PSU, altering within the known good range several voltage settings that WILL effect data integrity and taking a look at any software issues, RAID drivers etc is not haphazard. Kidding yourself and relying on a function that may not even report the correct data sounds so much better then?

    Sure, use SMART data logging, it's a useful tool, but it's not God.
    Coolermaster CM 690 II advance Case
    Corsair HX750 (CWT, 91%(80+ Gold rated @230V) single 62A 12V rail
    P55A-UD4 v2.0 @ F14
    Core i5 760 @ 20 x 201, 4.02GHz
    TRUE Black with a single Noctua NF-P12 pumping out 55 CFM @ 19db .
    2 x 2GB Mushkin Ridgeback (996902), @ 7-10-8-27, 2010-DDR, 1.66v
    2 x Gigabyte GTX 460 1024MB in SLI (Pre OC'd to 715MHz core and 1800MHz VRAM) @ 850 Core / 4100 Mem.
    Intel X25-M Boot Drive (OS and Programs) 200MB/s Read & 90MB/s Write
    Corsair X32 200MB/s Read & 100MB/s Write
    WD Caviar Blue 640GB C (Steam, Games, Storage, Temp Files & Folders, etc)
    Samsung F3 500GB Backup/Images
    Noctua 1300RPM 19dB case fan (rear extraction)
    3 x 140 MM Coolermaster LED fans (one front intake, one top extraction, one side intake)
    Dell Ultra Sharp 2209WAf E-IPS @ 1680x1050

  6. #26
    Join Date
    Jan 2010
    Location
    Someplace
    Posts
    17

    Default Re: Real problem, corruptions and disks inaccessable - eek! GA-EX58-UD5 board

    If you paid close attention to the google study. They made no mention of which manufacturer or model. In fact they quite specifically noted that it was a mix, but most data trends improved with specific models and series. Thus it is entirely possible to see a data skew based on a very small population of bad data providers. In fact, my own comparisons reveal highly different numbers at extremely large fleet sizes.

    Even they admit they admit that the data applicability varies greatly as time and technology changes.

    I requested they post the smart data so that I can help evaluate it. I actually did provide quick pointers on areas to examine. My tests would have taken five minutes to implement and a quick follow-up to examine the data. There were several steps provided in assisting to isolate the data.

    At the same time, I never discounted your methodology as completely. I simply said we should look at the data and make an informed decision. We are two sides of the same coin, but I'm championing analysis first.

    Not my problem if they get bored. I care as much as the person is willing to listen.

    In regarding failure scenarios as to possible reasons why.

    1) Poor power (consider the power source just as much as the cabling)
    2) Memory issues (Yes, memory issues can look just like file issues as least in regards to read errors)

    It's so cheap to mention SSDs and firmware. These are first and second generation devices by manufacturers quite new to the market. I would not be surprised to see smart failures. Hell they can't keep performance consistent.

    So in closing because I have other things to do. We are seeking the same thing. I would hope you take away some of the things I suggest and apply them to your methodology in the future. I will promise to read and absorb as much from your postings as well.

Page 3 of 3 FirstFirst 123

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •