Messages in this thread | | | Date | Mon, 25 Jan 1999 16:36:37 GMT | From | "Stephen C. Tweedie" <> | Subject | Re: Should raw I/O be added to the kernel? |
| |
Hi,
On 21 Jan 1999 23:38:15 -0500, uixjjji1@umail.furryterror.org (Zygo Blaxell) said:
>> Yes you can. The way these applications work is to write all of the
> No you can't. Suppose you send a bunch of raw writes to a SCSI disk > drive. OK, so the SCSI disk drive queues them in its embedded cache > RAM and tells the host CPU to send more data. Then the power fails > before the SCSI drive can flush its embedded cache.
> Oops.
Fine. If your disk hardware tells the host that it has completed an update to oxide and the update is still volatile, you have a broken disk: send it back. No enterprise-class databases support such hardware. For writeback caching to be supported, the cache _must_ be battery-backed or (in advanced multi-controller redundant storage cabinets) multipowered and multiported. That is a fundamental storage architecture issue, nothing at all to do with the host O/S.
> There's the possibility of external RAID devices that will undo all that > work for you by doing buffering and ACK's by themselves, then turning > around to talk to disks with data in cache.
Not on ANY decent storage systems.
--Stephen
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |