Now we come to the first major problem. These drives run on the Firewire bus system despite its own power supply with a large data throughput are not stable, they separate from the computer and this means due Raid 0 lengthy repair of data capture to data total loss. The first attempt at a solution was the interconnection of the two clusters for RAID 0 + 1 multi-project (RAID 01 with four plates [2 clusters] using the software Raid of OSX Mavericks).
Second problem: Good thought, but within a short time a huge problem: getting off of individual clusters on the Firewire bus led each to almost four days of recovery sessions: unacceptable! Especially because in the period of recovery so at any time may occur a second fault on the intact cluster, and that increases the risk of data loss total.
First solution and third problem: The hard drives on the Firewire must independently switched on and can be switched off, thus remains only a FireWire repeater / hub. I chose with my 3 Firewire disks / clusters to Delock 4 port IEEE 1394b FireWire Hub / Repeater. The independence of the storage media of one another was thus ensured. The problem with getting off the individual G-Technology Cluster remained.
The final solution for the time being: The stroke had to be provided with its own power supply (for he is also set). What are the minimum requirement for these power for stable operation, I can here not specify, because my solution was a pretty special: I use a battery backed up supply from old shortwave radio time, and now the RAID 0 + 1 system works from the two clusters initially properly.
Certain considerations should do just before buying in order to avoid the experienced disaster.
Conclusion:
- These clusters are clusters for fast video and image processing and the like. But here you are with this enormous capacity of 6 TB in the professional class.
- For pure non-mirrored data storage and backup media as they are unsuitable.
- In Firewire composite there is a risk that for any reason whatsoever "release" this cluster. Cascading on Firewirebus is completely unreliable. Here only operated over a reliably supplied FireWire Hub / Repeater.
- Who has two of these clusters, can produce a relatively safe system state over a FireWire repeater / hub with sufficient own power supply.
- For pure Daisychaining the Firewire bus, the cluster is unsuitable.
Was my mistake and I publish this review with the hope of others who have similar problems on Firewire help.
1 star trigger for the cluster with respect to the unreliability is justified.