Validate, Certify, Verify Failures

Verify Disk

A Verify Disk command is a simple read across the entire disk media. SoftRAID tells the disk to read all sectors and reports on its progress.

Verify Failure

A Verify Failure is very unusual. It means the disk can no longer be read at the error’s location. In most cases, the disk has failed. (always be alert to the possiblity of connection problems when you get any kind of disk error)

Validate Volume

Validate is the process of reading all the data in the volume, and comparing the data to the parity data, then updating the parity data.

The first time you validate a volume, you will see many blocks have updated parity. This is normal.

After you validate an HDD based volume once, you should never see additional parity blocks updated. This is because SoftRAID tracks all data written to the volume and calculates parity constantly.

Failed Validate

The two things that can cause a Validate failure are:

Open the SoftRAID log to get more information. If there is a read error on one of the disks, but not any others, this is likely a faulty disk and it should be replaced.

Certify Disk

Certification is a strenuous test of a disk. SoftRAID writes out a random pattern to the entire disk, then reads it back. Since the data written to the disk contains “checksums”, SoftRAID can confirm if the disk is reading back accurately.

It’s highly recommended to certify all new disks, as drives are sold untested. Certification helps eliminate bad drives before you put them into use. Be prepared that this certification proess takes time. One rotating media, over thunderbolt, a good estimate is one day for every 2TB of capacity. Using this estimate, a 14TB drive can take 7 days to certify.

SSD’s / NVMe drives are much faster, of course. A 4TB NVMe blade can complete a certify in less than a day.

Failed Certify

A disk should never fail a certify unless it is faulty, or there was a connection problem with the enclosure.

If your disk failed to “read” data during a certify, incorrect data was read back (All data written during a certify has a checksum attached so that SoftRAID can confirm the data was accurately read back as written), then it is likely that the disk is faulty.

Usually, a disk will fail or reallocate sectors when a disk is unable to read data.

A cetify may also fail if SoftRAID is unable to write to the disk or the disk “hangs”.

Example error message:

The disk (disk10) serial number (43eer32) on “Brandon’s Macbook Pro” failed to complete a certify.

Troubleshooting:

Note: If a disk loses connections to the host, you will get a read or write error. This is not the same as a read failure.

A read failure means the disk has failed to read and cannot reliably read the data at all on that section of the disk. Replace that disk immediately.

If the certify failed due to I/O errors, or read/write errors, and you are certain this was not caused by a power/connection problem, then the disk is unreliable and should not be used.

How to fix a Certify Failure:

Its recommended that you try again, as most certify failures are from connection issues and it is worth restarting the certification process.

Resume Certify:
Certify disk can be resumed.

Select the (disks) and “certify” again. Leave the same settings as before and, after authorizing with your password, you will be prompted to resume.