Tips For Replacing A Hard Drive From A Failed RAID

Tips For Replacing A Hard Drive From A Failed RAIDThere are some items to consider when replacing a hard drive from a failed RAID. If you are building a new RAID, then all hard drives in the array should be the identical model if at all possible. However, if you must replace a failed hard drive, it can sometimes be difficult to find the same model if that model is out of production.

Below are some tips to follow when selecting a replacement:

Keep in mind that the controller may or may not allow different models in a RAID, so check the RAID controller documentation.

Product life: What is the expected life of the remaining drives? If the other drives are approaching the end of their useful life, then it may be time to replace the entire RAID.

Capacity: The replacement drive should be the same or higher capacity than the original drive. Do not just look at the capacity on the box, since a few megabytes could make the difference between whether the drive will work or not.

(You should check the number of LBAs (or sectors) on the hard drive. Some RAID controllers will allow you to substitute larger drives if the exact capacity is not available, while other controllers require an exact match. Check with the controller manufacturer if the documentation doesn’t make it clear!)

Performance: The replacement drive should match the performance of the remaining drives as closely as possible. If your failed drive was 15,000 RPM, avoid replacing it with a 10,000 RPM drive. RAID arrays depend on the timing between drives to write data. Thus, if one drive doesn’t keep up, it may cause the entire array to fail or at least experience irritating problems.

Interface: Make sure the replacement drive uses the same type of interface connection as the failed drive. If the failed drive used a SCSI SCA (80-Pin) interface then don’t try to replace it with a 68-pin SCSI interface. With Seagate products the last two digits of the model number indicate the interface. For example: LW = 68-Pin, LC = 80-Pin.

The 80-pin LC drives are hot-swappable with backplane connections.

Cache Buffer: It is recommended that the cache buffer for each drive be the same value.  Most RAID controllers will consider drives with mismatching cache buffers to be ineligible for addition to a striped or parity array.

Read More

Tips for Replacing a Drive from a Failed RAID

Replace a drive from a Failed Raid There are several items to consider when replacing a drive from a failed RAID. If you are building a new RAID, then all drives in the array should be the identical model if at all possible.

However, if you must replace a failed drive, it can sometimes be difficult to find the same model if that model is out of production. Below are some tips to follow when selecting a replacement.

Note: Keep in mind that the controller may or may not allow different models in a RAID, so check the RAID controller documentation.

1. Product life: What is the expected life of the remaining drives? If the other drives are approaching the end of their useful life, then it may be time to replace the entire RAID.

2. Capacity: The replacement drive should be the same or higher capacity than the original drive. Do not just look at the capacity on the box, since a few megabytes could make the difference between whether the drive will work or not.

You should check the number of LBAs (or sectors) on the drive. Some RAID controllers will allow you to substitute larger drives if the exact capacity is not available, while other controllers require an exact match. Check with the controller manufacturer if the documentation doesn’t make it clear.

3. Performance: The replacement drive should match the performance of the remaining drives as closely as possible. If your failed drive was 15,000 RPM, avoid replacing it with a 10,000 RPM drive. RAID arrays depend on the timing between drives to write data. Thus, if one drive doesn’t keep up, it may cause the entire array to fail or at least experience irritating problems.

4. Interface: Make sure the replacement drive uses the same type of interface connection as the failed drive. If the failed drive used a SCSI SCA (80-Pin) interface then don’t try to replace it with a 68-pin SCSI interface. With Seagate products the last two digits of the model number indicate the interface.
For example:
LW = 68-Pin
LC = 80-Pin
The 80-pin LC drives are hot-swappable with backplane connections.

5. Cache Buffer: It is recommended that the cache buffer for each drive be the same value.  Most RAID controllers will consider drives with mismatching cache buffers to be ineligible for addition to a striped or parity array.

Read More