Florz

Member
  • Content Count

    20
  • Joined

  • Last visited

Community Reputation

0 Neutral

About Florz

  • Rank
    Member
  1. Florz

    Power Failure + SSD

    Enterprise class SSDs use (ultra/super-)capacitors for power backup as a standard feature. Unfortunately, these SSDs come at a much higher price point.
  2. One interesting point I failed to mention in my discussion above, is that SandForce SSD controllers include SandForce's RAID-like technology called RAISE (Redundant Array of Independent Silicon Elements). In theory, one should be able to recover all data even if one of the NAND flash chips is damaged. A SandForce based SSD may even continue to work with one failed NAND chip, if we are to believe the marketing hype. In addition to RAISE, SandForce controllers also employ technologies like deduplication/compression and AES-128 encryption. Various SSD manufacturers may also use customized firmwares with their own unique feature set. All these technologies and features add to the complexity of the recovery process. The raw data contained in a single NAND chip may appear to be completely random. You'll need a software tool that, in effect, emulates the particular SSD controller and firmware in order to extract the user data from the raw data.
  3. If all the individual NAND chips still work, it's possible. I believe equipment to recover individual NAND flash chips has already existed on the (professional) market for a while. Or you could make a NAND chip reader yourself if you have the necessary skills and resources. It would require desoldering all NAND chips from the PCB and extracting all the raw data with the chip reader. However, a successful recovery of all user data on an SSD would depend on the following factors: Detailed knowledge of the inner workings of the SSD controller (this may not be necessary if recovery of a particular SSD design is already supported in a commercial recovery solution). You'll probably not get this information from the manufacturer unless you are a large respectable company and sign the necessary NDAs. Then your only option is reverse engineering. A successful recovery of all metadata on the SSD. This is important since the data will be scattered in seemingly random order due to the wear-leveling mechanism. Without the metadata that contains the block address mapping tables it will be impossible to reconstruct the data. In the end this will probably not be feasible for the typical hardware hacker in the home basement, but I'm sure the leading recovery companies have already developed/acquired (or are in the process of developing/acquiring) the knowledge, skills and equipment to do these types of recoveries. But again, it all depends on the state of the individual NAND chips and if the metadata is recoverable.
  4. Ah, yes those are the regular SATA ports on your board supported by the Intel chipset. The two eSATA (= external SATA) ports connected to the JMicron controller are on the rear panel of the motherboard. The JMicron SATA/PATA controller is an extra chip on your motherboard that gives you the two eSATA ports and a PATA connector. The two eSATA ports are normally used for external drive enclosures/docks with eSATA connectors. Since you were talking about eSATA, I wasn't sure which ports you were using. You'll need an eSATA to SATA cable if you're going to try plugging the drive into one of these ports, but it won't be practical if the drive is inside the computer case. The six internal SATA ports are connected to the Intel ICH9R which is the southbridge part of the main chipset on the motherboard. Since you're using RAID, you've installed the Intel Matrix Storage Manager (or the newer Intel Rapid Storage Technology) driver, right? Which means the Intel driver (iastor.sys) is also handling the other internal SATA ports. The ICH9R is supposed to support RAID volumes up to 256TB, so I would think iastor.sys should be able to handle long LBA drives. But it's possible it isn't, or you're using an older driver version. You could try installing the latest Intel RST drivers. The latest version available on Intel's site is 9.6.4.1002. There's a leaked 9.6.5.1001 WHQL beta driver on this site. The Intel driver can also be used in AHCI mode. Be careful if you try these drivers on your RAID 0 install, messing with drivers when your boot drive is on a RAID 0 could leave you with an unbootable system. Backup your system first, or try them out on the test installation on the spare drive. You should also try Microsoft's own default drivers (msahci.sys in AHCI mode or intelide.sys in IDE mode) that comes with Win7. But that means abandoning your RAID 0 config. It's possible to replace the Intel RAID Option ROM in the BIOS with a newer version, but I can't say if that would help. You may find a modded BIOS for your motherboard around somewhere or mod it yourself, but modding the BIOS is a risky thing to do, you may end up with a dead motherboard. Since Win7 didn't detect the full capacity of the drive when you connected the drive after booting Win7, it's probably not a BIOS issue. You are able to boot now with the drive connected after you repartitioned the drive, right? Let's see how your tests go.
  5. The dock makes the drive look like a 4K sector drive with 732566646 sectors. But when you connect it directly to a SATA port, it's a 512-byte sector drive with 5860533168 sectors (or it emulates 512-byte sectors if it's an AF drive, but I'm not sure it's an AF drive). It's this large sector number that's causing trouble, since it requires more than 4 bytes to represent sector addresses (LBAs) this large. 746.39 GiB ~ 801 GB, indeed it does look like you've run into a 4 byte (32-bit) LBA limitation. Windows 7 should be long LBA-capable, so it's either a driver, controller or a BIOS issue. AFAICT your board has two eSATA-ports provided by a JMicron SATA controller with RAID as well as the six internal SATA ports provided by the intel ICH9R. Which controller are your boot drives connected to? Which controller did you connect the 3TB drive to? There are a few things you could try, like switching SATA modes, trying the other controller, trying different drivers, but the RAID 0 config limits your options a bit since you can't change the SATA mode of the controller you're currently running in RAID mode.
  6. Florz

    Best drive for WHS box

    The jumper workaround will only be effective for a single partition (the first). Since WHS uses two partitions - the system partition and a data partition - on the primary drive, it's probably best to avoid using an EARS drive as the primary drive in your WHS (v1), but they should be fine as secondary data drives with the jumper installed. This is what Microsoft says about advanced format drives and WHS: http://support.microsoft.com/kb/2385637/en-us
  7. Did you try adding all three storage driver modules to the modules file in your preferred order? And you'll need to run update-initramfs afterwards of course. That should work, at least according to this.
  8. Florz

    Is this drive dying?

    I looked up hostbyte 0x04 (DID_BAD_TARGET) from the error message. If I've understood correctly, it means that the target device was not found. Most likely the libata driver disabled the device after failing to communicate with it, ie. the drive became unresponsive. There should be more clues in the log (look for ata messages, soft/hard resets). That's probably also why smartctl failed. My advice is to check the cables and connections and run a diagnostic with the manufacturer's diagnostic utility. http://www.samsung.com/global/business/hdd/support/downloads/support_in_es.html
  9. Florz

    Discontinuing a RAID 0

    First you need to make sure what RAID type you're using. It's not possible to reset a RAID 0 without losing all data. In order to reuse the drives, the best thing to do is to backup all data and use the "Delete RAID Volume" menu option in the Intel Matrix Storage option ROM. If you're in fact using RAID 1, the "Reset Disks to Non-RAID" menu option in the Intel Matrix Storage option ROM will do what you want. Both drives may not always be bootable, but the data should be intact on both drives.
  10. Florz

    Is this drive dying?

    Try adding "-d ata" or "-d sat" to the smartctl command. # smartctl -a -d ata /dev/sde
  11. Florz

    Trouble upgrading disks in a RAID 1

    You'll have to revert to non-RAID, resize the first drive, and then create a new RAID-1 volume from the resized drive. Read page 4 of this document: http://download.intel.com/support/chipsets/imsm/sb/reference_content_intelmatrixstorageconsole.pdf From where you are now, the procedure will be something like this: Use the "Reset Disks to Non-RAID" menu option from the Intel Matrix Storage Manager option ROM. Turn off the system and disconnect the second drive (not strictly necessary if you make sure you don't get the drives mixed up in following steps). Boot into Windows from the first drive. Start your partition manager and resize the partition on the first drive. Turn off and reconnect the second drive. Boot into Windows (make sure you're still booting from the first drive). Open Disk Management and verify that C: is assigned to the first drive (the partition you resized earlier). Open the Intel Matrix Storage Console and click "Create RAID Volume from Existing Hard Drive" from the Actions menu. Follow the prompts and make sure you use the resized drive as the source drive. Good luck.
  12. Florz

    Problems formatting 2TB drive

    Interesting, it turns out it's a bug in Windows Server 2003 SP2 and a hotfix is available: http://support.microsoft.com/kb/979451
  13. Florz

    Win2K3 Server And Wd20Ears

    Installing the jumper has the effect of shifting the logical sector numbering by one internally by the drive, so that a partition starting at sector 63, is in fact starting at sector 64 instead. If you install the jumper you must leave the jumper on before and after partitioning and formatting. I would prefer to create an aligned partition instead of using the jumper. If the WD Align tool doesn't work with W2K3, you can create an aligned partition manually by following the procedure outlined in this Microsoft KB article: http://support.microsoft.com/default.aspx?scid=kb;EN-US;929491
  14. Florz

    Problems formatting 2TB drive

    I agree, it looks like a driver issue. Enabling SATA AHCI in BIOS and switching to AHCI drivers could solve the problem (if you're not already running in AHCI mode). Beware, however, enabling AHCI may cause Windows to BSOD during boot due to missing drivers. When troubleshooting this, you could probably save some time by doing a quick scan with HD Tune instead of doing a time consuming full format each time you're testing. Or you could use a disk hex editor (like the freeware HxD hexeditor) and see if you can read sectors >=2147483648 (0x80000000 hex) directly without getting errors.
  15. Florz

    How do you upgrade a RAID 1?

    You could also image your current RAID-1 volume using drive imaging software, remove the two RAID-1 drives, setup the new RAID-1 with the new drives, and finally restore the image to the new RAID-1. Most drive imaging software have the option to automatically resize the partition(s) to fit the new drive. But you'll need another drive large enough to hold the image file.