Date: Wed, 24 Mar 1999 10:37:17 +1030 From: Greg Lehey <grog@lemis.com> To: Codeguy <codeguy@wolf.com>, freebsd-questions@FreeBSD.ORG Subject: Re: Beginner questions re. vinum Message-ID: <19990324103717.K425@lemis.com> In-Reply-To: <199903232227.OAA19794@wolf.com>; from Codeguy on Tue, Mar 23, 1999 at 02:27:49PM -0800 References: <199903232227.OAA19794@wolf.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tuesday, 23 March 1999 at 14:27:49 -0800, Codeguy wrote:
> I'm setting up a new FreeBSD system (3.1) on a Pentium box with two
> Quantum Ultra-SCSI drives (4.5 GB each). This box will be used to
> log data from two important experiments. Since the data is kind of
> important, the powers that be want to have the data mirrored across
> the two drives.
>
> I've currently got 1.2 GB of each drive allocated; swap apace and
> root FS on da0, and /work on da1. I'd like to use the remaining 3.3
> GB of each drive for vinum volumes. Any reasons why I shouldn't be
> able to do so?
None that I can think of.
> I *have* read through the man pages (I've got hard copy here in
> front of me), and think I have an idea of how to proceed. I've
> created a file named /etc/vinum.rc, containing:
>
> drive drive0 device /dev/da0
> drive drive1 device /dev/da1
> volume www
> plex org striped 256k
> sd length 1600m drive drive0
> sd length 1600m drive drive1
> volume logs
> plex org striped 256k
> sd length 1600m drive drive0
> sd length 1600m drive drive1
Looks syntactically correct.
> As I understand what I've read, this should identify both hard
> drives as eligible for vinum control, identify volumes named "www"
> and "logs", each of which is built out of 256 kbyte stripes across
> both drives, for a total size of 1600 megs each volume. Is this
> more or less correct?
Partially. Each volume will have 3200 MB. And they won't be
mirrored. To mirror them, you need additional plexes. To judge from
what you say above, I think what you want is:
drive drive0 device /dev/da0
drive drive1 device /dev/da1
volume www setupstate
plex org striped 256k
sd length 800m drive drive0
sd length 800m drive drive1
plex org striped 256k
sd length 800m drive drive1
sd length 800m drive drive0
volume logs setupstate
plex org striped 256k
sd length 800m drive drive0
sd length 800m drive drive1
plex org striped 256k
sd length 800m drive drive1
sd length 800m drive drive0
This will give you mirrored, striped volumes of 1600 MB. Note that
the subdisks of the second plex are stored on the "other" disk
relative to the first plex. This means that any write operation to
the volume will go to both disks simultaneously. If you didn't do
this, the write to the second plex would have to wait for the write to
the first plex to complete.
> My next step was:
> vinum create /etc/vinum.rc
> which seemd to complete normally. This was followed by
> vinum init www
> vinum init logs
> which also seemed to complete normally (and very quickly as
> well - just a few seconds).
I suppose it depends on what you mean by "normally". I rather suspect
you got the message
www is not a plex
logs is not a plex
From vinum(8):
init plex
vinum init initializes a plex by writing zeroes to all its sub-
disks. This is the only way to ensure consistent data in a plex.
You must perform this initialization before using a RAID-5 plex.
It is also recommended for other new plexes.
> I then tried
> vinum start www
> vinum start logs
You don't need this unless the objects are down.
> which produced after just a few seconds:
> vinum: Can't write config top /dev/da0, error 19
> vinum: drive0 is down
> vinum: www.p0.s0 is crashed
> vinum: www.p0 is corrupt
> vinum: logs.p0.s0 is crashed
> vinum: logs.p0 is corrupt
> vinum: drive1 is down
> vinum: www.p0.s1 is crashed
> vinum: www.p0 is faulty
> vinum: www is down
> vinum: logs.p0.s1 is crashed
> vinum: logs.p0 is faulty
> vinum: logs is down
Interesting. I tried this here and got:
# vinum start obj
Can't start obj: Invalid argument (22)
I suppose we could do with a better message than that, but basically
it's trying to say "how can I start it when it's already started?".
I've changed the code, and now you'll get a message like:
# vinum start src.p0.s1
src.p0.s1 is already up
I've updated the man page, and soon you will be able to read (under
the description of the `start' command):
If object names are specified, vinum starts them. Normally this
operation is only of use with subdisks. The action depends on
the current state of the object:
o If the object is already in the up state, vinum does nothing.
o If the object is a subdisk in the down or reborn states,
vinum changes it to the up state.
o If the object is a subdisk in the empty state, the change de-
pends on the subdisk. If it is part of a plex which is part
of a volume which contains other plexes, vinum places the
subdisk in the reviving state and attempts to copy the data
from the volume. When the operation completes, the subdisk
is set into the up state. If it is part of a plex which is
part of a volume which contains no other plexes, or if it is
not part of a plex, vinum brings it into the up state immedi-
ately.
o If the object is a subdisk in the reviving state, vinum con-
tinues the revive operation offline. When the operation com-
pletes, the subdisk is set into the up state.
When a subdisk comes into the up state, vinum automatically
checks the state of any plex and volume to which it may belong
and changes their state where appropriate.
If the object is a volume or a plex, start currently has no ef-
fect: it checks the state of the subordinate subdisks (and plexes
in the case of a volume) and sets the state of the object accord-
ingly. In a later version, this operation will cause the sub-
disks
> And finally, once I have managed to complete the proper steps
> to actually have a working vinum config, what do I need to do
> to newfs my vinum volumes?
From vinum(4):
MAKING FILE SYSTEMS
Since vinum volumes do not contain partitions, the names do not need to
conform to the standard rules for naming disk partitions. For a physical
disk partition, the last letter of the device name specifies the parti-
tion identifier (a to h). vinum volumes need not conform to this conven-
tion, but if they do not, newfs will complain that it cannot determine
the partition. To solve this problem, use the -v flag to newfs.
To this I have added:
For ex-
ample, if you have a volume concat, use the following command to create a
ufs file system on it:
# newfs -v /dev/vinum/rconcat
As with other file systems, you perform the newfs command on the raw de-
vice, not the block device.
In addition to this, I'll write a section on how to set up vinum in
the vinum(8) man page. Real Soon Now.
Greg
--
When replying to this message, please copy the original recipients.
For more information, see http://www.lemis.com/questions.html
See complete headers for address, home page and phone numbers
finger grog@lemis.com for PGP public key
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990324103717.K425>