Date: Fri, 11 Feb 2005 13:05:13 +0100 From: Henning Kropp <wir02jcr@studserv.uni-leipzig.de> To: freebsd-stable@freebsd.org Subject: Re: (g)vinum with 5.3 Message-ID: <420C9F79.4090700@studserv.uni-leipzig.de> In-Reply-To: <200502091808.24679.msch@snafu.de> References: <4209E2FB.2070806@studserv.uni-leipzig.de> <200502091808.24679.msch@snafu.de>
next in thread | previous in thread | raw e-mail | index | archive | help
Matthias Schuendehuette schrieb:
>Hi Henning,
>
>Am Mittwoch, 9. Februar 2005 11:16 schrieb Henning Kropp:
>
>
>>[...]
>>The result I at boot time is a screen continuously printing out:
>>GEOM_VINUM: plex request failed for gvinum/plex/myvolume.p
>>
>>Does anybody know where to start from here?
>>
>>
>
>There are some points to check:
>
>a) Is this a RAID-5 Volume under 5.3-RELEASE? -> Upgrade to 5-STABLE!
>
>gvinum RAID-5 is known *not* to work under 5.3-RELEASE. The error was
>fixed during the freeze of the source tree for 5.3-RELEASE, so it was
>committed shortly ( 1 week or so) *after* the release was built.
>
>b) Check your vinum partitions with bsdlabel.
>
>In your case:
>
>bsdlabel /dev/ad0s1e
>bsdlabel /dev/ad1s1d
>bsdlabel /dev/ad2s1d
>
>Check if any of the vinum-type partitions have an offset of "0" (zero).
>
>
Thanks for your long response and explanation.
After hours (almost days building world takes awhile) of trying to fix
this problem I am facing a new situation.
I first upgraded to 5.3 and than tried to run (g)vinum! As I expected it
to work at least that way around. But it is not.
First of all my vinum setup I try to accomplish is a simple RAID-0:
drive a device /dev/ad0s1e
drive b device /dev/ad1s1c //has changed to s1c
drive c device /dev/ad2s1c / -"-
volume myvolume
plex org concat // I mist to post that the last time, sry!
sd length 3214325K drive a
sd length 3242353K drive b
sd length 3214243K drive c //size dont matter here
I checked bsdlabel as
root# bsdlabel /dev/ad1s1c
/dev/ad1s1c:
8 partitions:
size offset fstype [fsize bsize bps/cpg]
c: 78172227 63 4.2BSD 2048 16384 28552 # "raw" part,
don't edit
partition c: partition extends past end of unit
bsdlabel: partition c is not marked as unused!
bsdlabel: partition c doesn't start at 0!
bsdlabel: An incorrect partition c may cause problems for standard
system utilities
root# bsdlabel /dev/ad2s1c
# /dev/ad2s1c:
8 partitions:
size offset fstype [fsize bsize bps/cpg]
c: 78172227 63 4.2BSD 2048 16384 28552 # "raw" part,
don't edit
partition c: partition extends past end of unit
bsdlabel: partition c is not marked as unused!
bsdlabel: partition c doesn't start at 0!
bsdlabel: An incorrect partition c may cause problems for standard
system utilities
Both have an offset of 63, which is the default I think. But still
bsdlabel is not very happy with both. I sry but I cant tell why. If
anybody could I'll be happy to know.
But what happens when either try to run vinum or gvinum is:
vinum easily runs with this configuration but as the system reboots it
panics saying:
panic: umount. dangling vnode
I was told that this is to vfs_mount.c and geom_dev.c and a downgrade to
vfs_mount.c 1.27 and geom_de.c 1.75 would make the thing work again.
This is because the new versions together cant deal with every config,
for example it cant work with mine. How can I either downgrade or
rewrite my config?? (Maybe Mathias has written his PR)
Of course I tried to use gvinum right from the start. But with the
config gvinum tells me, that drive a (dev/ad0s1e, by now mounted with
/usr) is already known. Well, dont know what gvinum is tying to tell me
here.
I surely appreciate any help. Thanks!
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?420C9F79.4090700>