Re: [DRIVER SUBMISSION] DRBD wants to go mainline

From: Tomasz Chmielewski
Date: Sun Jul 22 2007 - 05:54:59 EST


Jan Engelhardt wrote

On Jul 22 2007 00:43, Lars Ellenberg wrote:

Think of it as RAID1 over TCP.

And what does it do better than raid1-over-NBD? (Which is already N-disk,
and, logically, seems to support cluster filesystems)

I don't know about DRDB, but NBD doesn't handle network disconnections at all (well, almost).

So basically, disconnect a NBD-connected system for a while (switch, cabling problem, operator error etc.), and you need lots of effort, perhaps restarts, to get the things to a functioning state (devices offlined, kicked out etc.).

I wouldn't call such raid-over-NBD setup reliable.


A better question would be: what does it do better than raid1-over-iSCSI?

iSCSI can recover from disconnections very well when configured properly; but when a disconnection is in place, most of the system will just "hang/freeze" (that is, from the user perspective - the system will be waiting for the I/O to complete, until the systems are connected again).


A brief reading of "official DRBD FAQ" didn't give me an answer to that problem.


--
Tomasz Chmielewski
http://wpkg.org
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/