XRaid Progress

In a previous post, I described our situation with storage and finding a way to reuse the XRaid's we had in the racks. By using a IDE2SATA adapter and a notebook 7200rpm drive, we were able to replace all the drives in a fairly cheap manner. Right now, I have 70 drives (5 XRaid's, 4 in picture.) formatting in Raid 5. We are about 24 hours in and formatting is at almost 50%.


  1. Hi Mike,
    I followed your guid since I was in the same situation. The only thing I did different is that I bought a different (very low cost ~$5) IDE2SATA controllers. Until the rebuilding of the array and formatting I did not notice any problems but as soon as I started to copy files on the array the xraid started to throw "Disk X Failed to respond. Retrying. RETRY COUNT:0 LBA:0x...". Lots of them continuously. This resulted in 1/2 minutes of halts in the copying. Then it would go again for a few minutes and then it would "freeze" again. No data loss but really annoying.

    So, I looked at your previous post again and I saw that the adapters you used are still available on-line.

    Q: How are your arrays going? Did you have any kind of problems of this sort? Or others?


  2. I ran into some issues where the drives would go offline. By removing the drive and re-adding it, it worked fine. We ended up going down a different route, it wasn't going to be fast enough for what I needed. Overall it worked pretty well with some minor quirks. The quirks were more seen when you start pushing large amounts of data through it.

  3. I've gone with this modification on the xServe Raid systems in my office... Using the Startech IDE to SATA converter and Hitachi 1 Terabyte 7200 RPM 2.5" drives.

    The overall performance is pretty decent and in fact seems faster than previously when on the OEM 750GB IDE/PATA drives. The added benefit of also running cooler since these SATA drives in their smaller form factor don't generate as much heat. I tried the 750GB Momentus Seagate hybrid drives... but I don't think these drives were fast enough and that the hybrid factor also caused the quirky behavior that you both described.

    I have it replicating to another xRaid sitting at a colocation and it seems to perform better with the modified sleds than the xRaid with the OEM 750GB drive at the colocation. So the next order of business on my plate is to convert the xRaid array at the colocation as well.