SmallNetBuilder

RAID - more

For the next test, I decided to add the remaining drive (drive 1) as a hot spare. From the RAID management page, I selected the RAID 5. Once selected, you can see in the box at the right and the icons across the bottom, that I had a number of options. I chose Configure Spare Drive.

Figure 11: RAID Management Tool showing three-drive RAID 5 and a single drive.

I had previously created several shares on drive 1. Through the Share Management utility, I could change the path for those shares to the RAID 5 volume, but in doing so, the data was not migrated. You have to do that manually. A nice feature would be to give the user the option of migrating the data when changing the location of the share. Once I selected Drive 1 as the spare and clicked on Configure spare drive, Drive 1 was added as a hot spare almost instantly.

Drive Fail Tests

You can probably see where this testing is headed. Now with the TS-409 configured for RAID 5 with a hot spare, I decided to "fail" one of the drives. I opened the front panel and removed drive 4. After about 15 seconds, the TS-409 beeped four times. 

Checking both the System Event Log and the RAID Management Tool, I could see that drive 4 had been removed, and that the RAID 5 was being rebuilt using the hot spare (drive 1). I also received an email to both configured addresses indicating that drive 4 had been removed. On the front panel, the LED for drive 4 went out, and the status LED blinked red, indicating RAID rebuild.  

During the rebuilding process, the shares were still available for read/write. This process took 3 hours and 30 minutes and completed successfully without data loss. When this process completed, I plugged in drive 4, and the system automatically designated it as the hot spare.

Figure 12: System Event Log showing re-mirroring, migrating to RAID 5, setting a hot spare and failing one of the drives in the RAID.

Figure 13: RAID 5 rebuilding onto drive 1, which was the hot spare before drive 4 was removed.

Since I know you'll ask, I also tested to see whether a RAID 6 configuration could really withstand two drive failures. You'll be happy to know that it can, although the process was not completely uneventful. I first reformatted and reinitialized the device for RAID6. Initialization didn't take too long, but a complete resync took a little over 9 hours. During that time, I was also copying 25GB of data to the device.

When the resync and copying was done, I pulled drives 2 and 3 out in rapid sequence. I received email notification only of the removal of drive 2, perhaps because I didn't wait for the device to beep between drive removals. Still, the management utility showed both drives 2 and 3 removed and that the device was operating in a degraded mode. However, even with two drives removed, I still had access to the shares.

Next, I re-inserted drive 2 and 3, again one right after the other. The log showed only drive 2 being plugged in, but the drive management utility again showed both drive 2 and drive 3. Drive 3, however, showed 0 MB available.

The RAID array first successfully added drive 2 (the first on reinserted) back into the array (including another 9 hour resync). But drive 3, the second drive inserted, continued to show 0 MB available and a "scan drive" on that drive failed. It wasn't until I rebooted, that drive 3 was recognized, and added back into the RAID 6 array (including yet another 9 hour resync).