headers
t.blum | 2 Oct 2006 12:52
Picon
Favicon

FreeBSD 6.3 Raid1 (geom) configuration problem with multiple partitions

Hi there,

I recently switched using FreeBSD on my server (used Linux before). I've a problem with configuring geom as
raid1 for my two identical hd's. I started with installing a FreeBSD 6.3 on my ad0 disk. The install uses 1
slice with the default FreeBSD partition table:

ad0s1a (root) 
ad0s1b (swap) 
ad0s1d (var) 
ad0s1e (tmp) 
ad0s1f (usr)

Everything is working and at this point I started configuring geom with help of this:
http://people.freebsd.org/~rse/mirror/.  tutorial (approach1). The following commands worked like
a charm (second disk is ad3):

dd if=/dev/zero of=/dev/ad3 bs=512 count=79 
gmirror label -v -n -b round-robin gm0 /dev/ad3 
gmirror load fdisk -v -B -I /dev/mirror/gm0 
bsdlabel -w -B /dev/mirror/gm0s1 
bsdlabel -e /dev/mirror/gm0s1 

But when I look at /dev/mirror/ it only created 2 partitions, gm0s1a and gm0s1c so the following commands:

newfs -U /dev/mirror/gm0s1* 
mount /dev/mirror/gm0s1* /mnt/*

Will not work for gm0s1d (var) gm0s1e (tmp) and gm0s1f (usr). Does anybody know what I am doing wrong?

ps. Please be gentle, i'm just a newbie ;)
(Continue reading)

Permalink | Reply | 2 comments |
headers
FreeBSD bugmaster | 2 Oct 2006 13:08
Picon
Favicon

Current problem reports assigned to you

Current FreeBSD problem reports
Critical problems
Serious problems

S Tracker      Resp.      Description
--------------------------------------------------------------------------------
o kern/73177   geom       kldload geom_* causes panic due to memory exhaustion
o kern/76538   geom       [gbde] nfs-write on gbde partition stalls and continue
o kern/84556   geom       [geom] GBDE-encrypted swap causes panic at shutdown
o kern/87544   geom       [gbde] mmaping large files on a gbde filesystem deadlo
o kern/89102   geom       [geom_vfs] [panic] panic when forced unmount FS from u
o bin/90093    geom       fdisk(8) incapable of altering in-core geometry
o kern/90582   geom       [geom_mirror] [panic] Restore cause panic string (ffs_
o kern/98034   geom       [geom] dereference of NULL pointer in acd_geom_detach 

8 problems total.

Non-critical problems

S Tracker      Resp.      Description
--------------------------------------------------------------------------------
o bin/78131    geom       gbde "destroy" not working.
o kern/79251   geom       [2TB] newfs fails on 2.6TB gbde device
o kern/94632   geom       [geom] Kernel output resets input while GELI asks for 

3 problems total.
Permalink | Reply | 0 comments |
headers
Eric Anderson | 2 Oct 2006 13:40
Favicon

Re: FreeBSD 6.3 Raid1 (geom) configuration problem with multiple partitions

On 10/02/06 05:52, t.blum <at> home.nl wrote:
> Hi there,
> 
> I recently switched using FreeBSD on my server (used Linux before). I've a problem with configuring geom
as raid1 for my two identical hd's. I started with installing a FreeBSD 6.3 on my ad0 disk. The install uses 1
slice with the default FreeBSD partition table:
> 
> ad0s1a (root) 
> ad0s1b (swap) 
> ad0s1d (var) 
> ad0s1e (tmp) 
> ad0s1f (usr)
> 
> Everything is working and at this point I started configuring geom with help of this:
http://people.freebsd.org/~rse/mirror/.  tutorial (approach1). The following commands worked like
a charm (second disk is ad3):
> 
> dd if=/dev/zero of=/dev/ad3 bs=512 count=79 
> gmirror label -v -n -b round-robin gm0 /dev/ad3 

So far, you've created a mirror (gm0) with one device in it (ad3).

> gmirror load fdisk -v -B -I /dev/mirror/gm0 

I think you meant:

gmirror load
fdisk -v -B -I /dev/mirror/gm0

> bsdlabel -w -B /dev/mirror/gm0s1
(Continue reading)

Permalink | Reply | 1 comment |
headers
Daniel Bristot de Oliveira | 2 Oct 2006 17:53
Picon

Re: FreeBSD 6.3 Raid1 (geom) configuration problem with multiple partitions

2006/10/2, Eric Anderson <anderson <at> centtech.com>:
> On 10/02/06 05:52, t.blum <at> home.nl wrote:
> > Hi there,
> >
> > I recently switched using FreeBSD on my server (used Linux before). I've a problem with configuring geom
as raid1 for my two identical hd's. I started with installing a FreeBSD 6.3 on my ad0 disk. The install uses 1
slice with the default FreeBSD partition table:
> >
> > ad0s1a (root)
> > ad0s1b (swap)
> > ad0s1d (var)
> > ad0s1e (tmp)
> > ad0s1f (usr)
> >
> > Everything is working and at this point I started configuring geom with help of this:
http://people.freebsd.org/~rse/mirror/.  tutorial (approach1). The following commands worked like
a charm (second disk is ad3):
> >
> > dd if=/dev/zero of=/dev/ad3 bs=512 count=79
> > gmirror label -v -n -b round-robin gm0 /dev/ad3
>
> So far, you've created a mirror (gm0) with one device in it (ad3).
>
> > gmirror load fdisk -v -B -I /dev/mirror/gm0
>
> I think you meant:
>
> gmirror load
> fdisk -v -B -I /dev/mirror/gm0
>
(Continue reading)

Permalink | Reply | 0 comments |
headers
Jeremie Le Hen | 3 Oct 2006 18:22

Weird thing with gmirror

Hi,

(Please Cc: me in your replies, I am not subscribed.)

I am using:
% obiwan:tataz$ uname -mv
% FreeBSD 6.1-STABLE #8: Sun Jun 25 18:27:14 CEST 2006     root <at> obiwan:/usr/obj/usr/src/sys/OBIWAN  i386

I have set up a mirror gm1 using gmirror(8), with ad1 and ad2 as providers.
This had worked pretty well for weeks, until I got the following error:

% ad1: FAILURE - READ_DMA status=51<READY,DSC,ERROR> error=40<UNCORRECTABLE> LBA=6170832
% GEOM_MIRROR: Request failed (error=5). ad1[READ(offset=3159465984, length=65536)]
% GEOM_MIRROR: Device gm1: provider ad1 disconnected.

Obviously, my hard drive (Maxtor's drives really, really suck) is going
sick.  But as usual, a mere reboot seems to improves the disk state, for
a while at least.

I have rebooted the box, but now the kernel says:

% GEOM_MIRROR: Device gm1 created (id=1340871616).
% GEOM_MIRROR: Device gm1: provider ad1 detected.
% GEOM_MIRROR: Device gm1: provider ad2 detected.
% GEOM_MIRROR: Component ad1 (device gm1) broken, skipping.

And gmirror(8)'s output is:

% Geom name: gm1
% State: DEGRADED
(Continue reading)

Permalink | Reply | 1 comment |
headers
Damian Wiest | 3 Oct 2006 20:40
Favicon

Re: Weird thing with gmirror

On Tue, Oct 03, 2006 at 06:22:16PM +0200, Jeremie Le Hen wrote:
> Hi,
> 
> (Please Cc: me in your replies, I am not subscribed.)
> 
> I am using:
> % obiwan:tataz$ uname -mv
> % FreeBSD 6.1-STABLE #8: Sun Jun 25 18:27:14 CEST 2006     root <at> obiwan:/usr/obj/usr/src/sys/OBIWAN  i386
> 
> I have set up a mirror gm1 using gmirror(8), with ad1 and ad2 as providers.
> This had worked pretty well for weeks, until I got the following error:
> 
> % ad1: FAILURE - READ_DMA status=51<READY,DSC,ERROR> error=40<UNCORRECTABLE> LBA=6170832
> % GEOM_MIRROR: Request failed (error=5). ad1[READ(offset=3159465984, length=65536)]
> % GEOM_MIRROR: Device gm1: provider ad1 disconnected.

One of the drives had a read error and was removed from the mirror.

> Obviously, my hard drive (Maxtor's drives really, really suck) is going
> sick.  But as usual, a mere reboot seems to improves the disk state, for
> a while at least.

There's no need to reboot.

> I have rebooted the box, but now the kernel says:
> 
> % GEOM_MIRROR: Device gm1 created (id=1340871616).
> % GEOM_MIRROR: Device gm1: provider ad1 detected.
> % GEOM_MIRROR: Device gm1: provider ad2 detected.
> % GEOM_MIRROR: Component ad1 (device gm1) broken, skipping.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Ivan Voras | 3 Oct 2006 22:32
Picon

Re: Announcing: gvirstor

Yoshihiro Ota wrote:

> These are the command lines I tried
> # mount /dev/virstor/test /mnt/tmp
> # tar xf ports.tar.bz -C /mnt/tmp &
> # ./gvirstor remove test md3
> # ./gvirstor add test md4
> # ./gvirstor status test
> # ./gvirstor add test md3
> 
> Kernel paniced after a while. Many things were going so it was not sure that gvirstor was the cause of the problem.

To revisit this thread - I tried steps similar to those above (using
real devices instead of md) and can't reproduce the panic.
Permalink | Reply | 0 comments |
headers
Johan Ström | 7 Oct 2006 23:13

gmirror, removing disk and mounting gives strange error

Hi

I just gmirrod (gm1) two drives (ad8 and ad12) on a box. Working  
great and all. Now I needed some extra space while moving around  
files, and I removed ad8:

gmirror remove gm1 ad8

Working fine, gmirror reporting running complete on ad12 only. nice.
Now I try to mount one of the partitions on ad8 (ad8s1g):

crus /# mount /dev/ad8s1g /mnt
mount: /dev/ad8s1g: Operation not permitted

Huh.. wtf.. ok, so I check dmesg:

WARNING: R/W mount of /storage/gm1s1g denied.  Filesystem is not  
clean - run fsck

gm1?? I tought it was removed from gm1..

# gmirror dump /dev/ad8
Can't read metadata from /dev/ad8: Invalid argument.

dumping ad12 for example gives me correct info.. so it seems ad8 is  
cleared correctly.. trying to run clear on ad8 gives same errors as  
above.

Okay, so the FS might not be clean since I removed the disk while the  
FS was mounted.. But why does it report gm1s1g?? thats the mountpoint
(Continue reading)

Permalink | Reply | 4 comments |
headers
R. B. Riddick | 7 Oct 2006 23:30
Picon
Favicon

Re: gmirror, removing disk and mounting gives strange error

--- Johan Ström <johan <at> stromnet.org> wrote:
> crus /# mount /dev/ad8s1g /mnt
> mount: /dev/ad8s1g: Operation not permitted
> 
> Huh.. wtf.. ok, so I check dmesg:
> 
> WARNING: R/W mount of /storage/gm1s1g denied.  Filesystem is not  
> clean - run fsck
> 
R u sure, that the message in dmesg is caused by ur mount attempt?

If u try again to mount it, do u see another message in dmesg or in
/var/log/messages?

-Arne

---

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com
Permalink | Reply | 1 comment |
headers
Ivan Voras | 8 Oct 2006 00:23
Picon

Re: gmirror, removing disk and mounting gives strange error

Johan Ström wrote:

> Okay, so the FS might not be clean since I removed the disk while the FS
> was mounted.. But why does it report gm1s1g?? thats the mountpoint of my
> gm1 mirror (or rather the s1g on the gm1 mirror).. Confusing ;)

UFS records the mountpoint where the file system was last mounted. I've
seen fsck use this information when reporting error messages, so this
may be the reason.
Permalink | Reply | 1 comment |

Gmane