Before leaving this discussion of features, I’d like to mention a couple of features, which, while not new on the Quad, nevertheless, I find quite useful. First there’s Web Access which I covered in the review of the LinkStation Duo. It’s quite simple to set up, and once enabled, you have access to your files via a web browser. There’s no complicated DDNS setup nor is there a third party involved as with Western Digital’s MioNet.
Second, I like the built-in Disk Backup feature. With this feature enabled, you can back up any Buffalo NAS to another Buffalo NAS on your network. You can schedule up to 10 backup jobs or run a job immediately. You can also back up to a USB drive directly attached to the NAS.
Drive Fail Test
One of the reasons you buy a four drive system and run it in a fault-tolerant configuration is to ensure that you don’t lose any data. To see how the LinkStation Live responded to a drive failure, I pulled one of the drives (drive 2) out. Here’s what I found:
- Within a few seconds the status light started blinking.
- The Quad sent out a series of emails (10 total) over the course of about 45 seconds
- Next, the Quad shut down!
Apparently, one of the configuration options, set by default, is for the NAS to shut down when it detects a RAID failure. I would think that the default option should be to notify, but leave the device online to continue to service client requests in the degraded mode. For small offices that rely on outsourced IT support, that would leave the NAS available until a technician could arrive and assess the problem.
I powered the NAS back on and it booted into degraded mode. There were three levels of notification.
- The status light flashed in sequences of four flashes, which according to the user guide, indicates a RAID problem. The LED for drive two (the one I pulled) was now red. Normal status is green.
- The home page of the web admin page indicated that there was a RAID problem. (“Error has occurred in RAID Array 1”)
- NASNavigator (Figure 12) showed the error condition. With my mouse hovering over the icon, additional information appeared indicating which drive was missing and that the device was operating in a degraded mode.
Figure 12: NASNavigator showing RAID error condition
What really concerned me, however, was the wording in some of the emails that indicated that I may have lost data. Nor were the emails particularly helpful in letting me know exactly what had happened. I wasn’t able to confirm if I did, in fact, lose data, but the whole point of a RAID 5 configuration is to safeguard against a data loss. Here are a couple of the email notifications I received:
From: "LS-QL812 (LinkStation)" <firstname.lastname@example.org> To: <email@example.com> Sent: Sunday, December 14, 2008 5:14 PM Subject: LinkStation Status Report:DISK Error Notification: Error:Write error > DISK Error Notification > > HDD error occurred > Disk(s) the error occurred:UnknownDisk() > > (sdb) WRITE sector:12016445 count:6 > Disk writing error > Some data may not be recoverable. > Immediate change of disk is recommended. > > > [LinkStationInformation] > LinkStationName: LS-QL812 (LS-QL) > Time: 2008/12/14 17:14:42 > Setting Screen: http://192.168.100.181/ From: "LS-QL812 (LinkStation)" <firstname.lastname@example.org> To: <email@example.com>Sent: Sunday, December 14, 2008 5:15 PM Subject: LinkStation Status Report:RAID Error Notification: Error:RAID de-grade occurred (RAID Array 1 (md2)) > RAID Error Notification > > Target array:RAID Array 1 (md2) > Disk(s) the error occurred:UnknownDisk(sdb) > > RAID array entered de-grade mode. > > [LinkStationInformation] > LinkStationName: LS-QL812 (LS-QL) > Time: 2008/12/14 17:15:09 > Setting Screen: http://192.168.100.181/
The instruction manual clearly states that the drives in the LinkStation Quad are not hot swappable – something I also confirmed with their support engineer. However, Buffalo’s web site shows this contradictory (and incorrect) information:
In addition to its RAID drive security, the LinkStation Quad has a snap-off front plate, which gives easy access to the quick-swap hard drives, which allow for a drive replacement without the need for shutting down the unit.
Not wanting to take a chance on damaging the drive, I followed the instruction manual and shut down the Quad and unplugged the power supply. I then re-inserted the “Failed” drive to see if the LinkStation Quad would automatically start to rebuild the array.
Apparently that’s not an automatic process – at least for the failure mode I created. When the Quad came back up, I still got the error notification on the web admin home page. The instruction manual didn’t really tell me how to proceed, so I went to the disk management page, selected the drive that I had just inserted (drive 2) and clicked on “Restructure Array”. (See Figure 13)
Figure 13: Restructure RAID Array
After acknowledging that all data on drive 2 would be wiped out and then entering a four-digit confirmation number, the array restructure began. During the array restructuring, file sharing is disabled, as is the ability to make any further configuration changes. After about 15 minutes or so, the restructuring of the array completed, and the RAID “repair”, which is actually are rebuild of the array, began with an estimated time of 232 minutes. E-mail notifications confirmed that the RAID repair took 3 hours and 36 minutes. During the repair process, I once again had access to file sharing and could change settings.
My experience was very similar to the RAID rebuild process I experienced when I tested Buffalo’s TeraStation Pro II last year. A conversation with a Buffalo rep. indicated that for a new, unformatted drive, such as Buffalo would supply to replace a defective drive under warranty, the rebuild process would start automatically. The manual process that I went through is intended to prevent people from accidentally erasing data from a drive that’s already been initialized.
I also went back and reconfigured the option to not shut down on RAID failure, and again, pulled a drive – this time while I was writing data to the drive. The Quad did revert into “degraded mode”, but the files continued to be copied to the device. I did, however, continue to receive email notifications about write errors even after the Quad reported that it was operating in a degraded mode. Windows didn’t report any problems with the files copied, and the file, directory count and number of bytes matched when I compared the source and destination of files copied.
This mode wasn't error-free, however. Even though the LED on the failed drive (drive 3 for this test) turned red, the drive properties information under both Disk Management and System Status still showed the original drive properties. However, the RAID structure showed no check mark next to the failed drive. I then shut down the LinkStation Quad and rebuilt it as per my original test.