We use 3ware Inc 9650SE SATA-II RAID PCIe RAID controllers with battery backed write cache. Our spare hardware has the same controller.

I was wondering; are these controllers smart enough not to sync the cache when the disks have been changed? For example, if I deploy one of those spare machines by putting in the disks of another machine and that spare machine still has pending writes, will it be smart enough not to perform those writes on the replaced array?

Edit: my scenario is not really made clear, so let me give an example:

  1. server1 goes down because of power supply failure.
  2. I put the disks in server2 and start.
  3. I repair server1
  4. I put the disks back from server2 in server1 (it's not relevant right now that in reality I would probably keep server2 running).
  5. If server1 doesn't have safeguards, it will write to the array, thinking it's simply powering up again, corrupting it.

closed as too localized by Greg Askew, mdpc, Scott Pack, Magellan, Michael Hampton Dec 6 '12 at 23:17

This question is unlikely to help any future visitors; it is only relevant to a small geographic area, a specific moment in time, or an extraordinarily narrow situation that is not generally applicable to the worldwide audience of the internet. For help making this question more broadly applicable, visit the help center. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    Why would the spare server have pending or uncommitted writes in its cache? – ewwhite Jun 4 '12 at 13:20
  • I doubt it will try to flush to new drives, but if the spare hardware has been powered off for a long enough time then it will not have any charge left (and thus no spare data). My old workstation card was a 9750 and the BBU only kept data for a few days. (Granted, a 9650SE is not the same as a 9750SE, but close) – Hennes Jun 4 '12 at 13:22
  • But what when the battery would not be empty? – Halfgaar Jun 4 '12 at 16:39

If server1 which has battery backed write cache has power supply failure, taking it down, you should move your battery backed cache with battery connected to server2 to ensure that the cached content is written to the disks, so the scenario you mention should not happen normally.

  • Stupid question maybe, but the cache is not separate from the controller, right? – Halfgaar Jun 4 '12 at 19:58

If you have pending writes in cache and you take the disks out from under it, those disks will be corrupt. If you then put new disks in, I don't know what would happen, but even if it wrote the pending IO, it would be gibberish and you'd be formatting the new drives anyways.

Not the answer you're looking for? Browse other questions tagged or ask your own question.