Bad Hard Drive taking whole server down - question

[Frank Broughton]

I beg to ask, how can a bad hard disk take a server completely offline? Is that not what RAID is for? Are we lacking the proper monitoring on the servers at JPC to notice bad drive before a complete server is down? Just asking that is all. If I am wrong with this thinking please let me know.

[JPC-NickO]

Hi Frank,

The idea behind raid is to allow the server to keep going during a disk failure. The problem is that the RAID cards are not perfect (they are really just a specialized computer that is designed to manage data). Sometimes when a drive fails it's controller (the embedded controller on the disk) sends screwy messages to the RAID controller and sometimes this is handled badly (read: the controller locks up or does something that the OS/Driver doesn't like) which results in the entire system becoming unstable. These aren't cheap controllers we are using either -- they cost around $300-500 a controller (without drives).

You should note that for every drive failure that you hear about on the forums there are 5-10+ dealt seamlessly and without any service interruptions. We only put up notices when there will be outages or performance issues.

[Frank Broughton]

Thanks Nick. I guess they take a long time to rebuild? Been 4 hours since the RAID rebuild started. Next time that hardware decides to fail make sure it happens on any day but a Sunday okay - hahaha I have a couple churches on the server that live stream their Sunday services. They will understand though. Just another thing to deal with and when many are matters of life, a down server is not high on the list. Thanks.... Hope this rebuild does the trick! If you look bad at the records SNAKE had a bad start! Snake was being snakie... Maybe church websites should not be on a server called snake anyways... haha