October 3, 201213 yr My test-server is built of the following components: SM X9SCM-F + E3-1240 + 16 GB ECC RAM Areca ARC1200 + 2 2TB Seagates M1015/LSI2008 + 2 x 1 TB Samsung 1x 750 GB Samsung in 2 SM CSE-M35T1. Seasonic X760 The last couple of weeks it happens that a disk red-balls but SMART reports indicate that the disks are ok. The disk is re-assigned to the array, rebuilt, parity checked and everything is ok for a couple of days, then it starts over again. How to find the root-cause of why this happens and how can the reliability be tested of all this?
October 3, 201213 yr If a disk is disabled, then a "write" to it failed. It could be the power connection to the drive. It could be the power cable/splitter/backplane/drive-cage connection to the drive. It could be the sata connection to the drive, or sata cable, or backplane/drive-cage connection to the drive. It could be the disk controller, or anything connected to the same disk controller that might cause it to lock up. (a disk on a different channel??) It could be a poorly regulated power supply, unable to handle the peak demands of all your disks. It could be the disk itself, sensitive to vibration, or heat, temperature-changes, or voltage. A SMART test just shows how the disk looks as of the moment it is run. It will not tell you every time the back door to your home is slammed shut, the disk vibrates in the drive tray. (theoretical example), or loose cable loses connection (another possibility)
October 3, 201213 yr Author Thanks Joe, how will a failed write show up during a preclear? Or won't it show up at all?
Archived
This topic is now archived and is closed to further replies.