ccd and gstripe on 5.3-RC

From: Miguel Mendez <flynn_at_energyhq.es.eu.org>
Date: Fri, 22 Oct 2004 18:58:25 +0200
Hi,

I installed 5.3-RC1 yesterday. My main box has 2 9GB disks, which I
usually partition this way:

da0:[512MB for /       ][ 8.5GB ]
da1:[512MB for swap][ 8.5GB ]

I combine the two 8.5GB partitions (/dev/da[01]s1d) in a stripped
volume. I did it with vinum in the past, and have done it with ccd for a
while.

However, after the install I discovered the gstripe command, and assumed
(because it's not mentioned otherwise in the man page) that it would
work with partitions like ccd/RAIDframe. So I followed the example and
create an automatic stripe called usr using the two 'd' partitions. It
worked fine (read: gstripe didn't complain). I even newfs'd and mounted
it, and proceeded to copy the original /usr to the stripped volume.
Everything went fine, so I rebooted the box. And then gstripe started
printing a lot of errors about da0 and da1, attached the volume and then
destroyed it. Sorry, I don't have the log because I reinstalled after
that. My question is, is gstripe supposed to work *only* with whole
disks? If that's the case, wouldn't it be interesting to make it be able
to work with partitions as well?

In the end, I did my usual ccd.conf install, and that's the second part
of the message. The man page for ccdconfig ends with this sentence:

"If you need more than this you should look into external hardware
RAID SCSI boxes, RAID controllers (see GENERIC), or software RAID
systems such as RAIDframe raid(4) and vinum(8)."

However, raid(4) doesn't exist anymore (in RELENG_5, that is), shall I
submit a patch for that one (the change is trivial) ?

Cheers,
-- 
	Miguel Mendez <flynn_at_energyhq.es.eu.org>
	http://www.energyhq.es.eu.org
	PGP Key: 0xDC8514F1
	Note: All HTML and non-english mail goes to /dev/null

Received on Fri Oct 22 2004 - 14:57:09 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:19 UTC