freebsd-nq/share/man/man4/vinum.4

1023 lines
34 KiB
Groff

.\" Hey, Emacs, edit this file in -*- nroff-fill -*- mode
.\"-
.\" Copyright (c) 1997, 1998
.\" Nan Yang Computer Services Limited. All rights reserved.
.\"
.\" This software is distributed under the so-called ``Berkeley
.\" License'':
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\" 3. All advertising materials mentioning features or use of this software
.\" must display the following acknowledgement:
.\" This product includes software developed by Nan Yang Computer
.\" Services Limited.
.\" 4. Neither the name of the Company nor the names of its contributors
.\" may be used to endorse or promote products derived from this software
.\" without specific prior written permission.
.\"
.\" This software is provided ``as is'', and any express or implied
.\" warranties, including, but not limited to, the implied warranties of
.\" merchantability and fitness for a particular purpose are disclaimed.
.\" In no event shall the company or contributors be liable for any
.\" direct, indirect, incidental, special, exemplary, or consequential
.\" damages (including, but not limited to, procurement of substitute
.\" goods or services; loss of use, data, or profits; or business
.\" interruption) however caused and on any theory of liability, whether
.\" in contract, strict liability, or tort (including negligence or
.\" otherwise) arising in any way out of the use of this software, even if
.\" advised of the possibility of such damage.
.\"
.\" $FreeBSD$
.\"
.Dd October 5, 1999
.Dt vinum 4
.Os FreeBSD
.Sh NAME
.Nm vinum
.Nd Logical Volume Manager
.Sh SYNOPSIS
.Cd "kldload vinum"
.Cd "kldload Vinum"
.Sh DESCRIPTION
.Nm
is a logical volume manager inspired by, but not derived from, the Veritas
Volume Manager. It provides the following features:
.Bl -bullet
.It
It provides device-independent logical disks, called \fIvolumes\fP. Volumes are
not restricted to the size of any disk on the system.
.It
The volumes consist of one or more \fIplexes\fP, each of which contain the
entire address space of a volume. This represents an implementation of RAID-1
(mirroring). Multiple plexes can also be used for
.\" XXX What about sparse plexes? Do we want them?
.if t .sp
.Bl -bullet
.It
Increased read throughput.
.Nm
will read data from the least active disk, so if a volume has plexes on multiple
disks, more data can be read in parallel.
.Nm
reads data from only one plex, but it writes data to all plexes.
.It
Increased reliability. By storing plexes on different disks, data will remain
available even if one of the plexes becomes unavailable. In comparison with a
RAID-5 plex (see below), using multiple plexes requires more storage space, but
gives better performance, particularly in the case of a drive failure.
.It
Additional plexes can be used for on-line data reorganization. By attaching an
additional plex and subsequently detaching one of the older plexes, data can be
moved on-line without compromising access.
.It
An additional plex can be used to obtain a consistent dump of a file system. By
attaching an additional plex and detaching at a specific time, the detached plex
becomes an accurate snapshot of the file system at the time of detachment.
.\" Make sure to flush!
.El
.It
Each plex consists of one or more logical disk slices, called \fIsubdisks\fP.
Subdisks are defined as a contiguous block of physical disk storage. A plex may
consist of any reasonable number of subdisks (in other words, the real limit is
not the number, but other factors, such as memory and performance, associated
with maintaining a large number of subdisks).
.It
A number of mappings between subdisks and plexes are available:
.Bl -bullet
.It
\fIConcatenated plexes\fP\| consist of one or more subdisks, each of which
is mapped to a contiguous part of the plex address space.
.It
\fIStriped plexes\fP\| consist of two or more subdisks of equal size. The file
address space is mapped in \fIstripes\fP, integral fractions of the subdisk
size. Consecutive plex address space is mapped to stripes in each subdisk in
.if n turn.
.if t \{\
turn.
.ig
.\" FIXME
.br
.ne 1.5i
.PS
move right 2i
down
SD0: box
SD1: box
SD2: box
"plex 0" at SD0.n+(0,.2)
"subdisk 0" rjust at SD0.w-(.2,0)
"subdisk 1" rjust at SD1.w-(.2,0)
"subdisk 2" rjust at SD2.w-(.2,0)
.PE
..
.\}
The subdisks of a striped plex must all be the same size.
.It
\fIRAID-5 plexes\fP\| require at least three equal-sized subdisks. They
resemble striped plexes, except that in each stripe, one subdisk stores parity
information. This subdisk changes in each stripe: in the first stripe, it is the
first subdisk, in the second it is the second subdisk, etc. In the event of a
single disk failure,
.Nm
will recover the data based on the information stored on the remaining subdisks.
This mapping is particularly suited to read-intensive access. The subdisks of a
RAID-5 plex must all be the same size.
.\" Make sure to flush!
.El
.It
.Nm Drives
are the lowest level of the storage hierarchy. They represent disk special
devices.
.It
.Nm
offers automatic startup. Unlike UNIX file systems,
.Nm
volumes contain all the configuration information needed to ensure that they are
started correctly when the subsystem is enabled. This is also a significant
advantage over the Veritas\(tm File System. This feature regards the presence
of the volumes. It does not mean that the volumes will be mounted
automatically, since the standard startup procedures with
.Pa /etc/fstab
perform this function.
.El
.Sh KERNEL CONFIGURATION
.Nm
is currently supplied as a kernel loadable module (kld), and does not require
configuration. As with other klds, it is absolutely necessary to match the kld
to the version of the operating system. Failure to do so will cause
.Nm
to issue an error message and terminate.
.Pp
It is possible to configure
.Nm
in the kernel, but this is not recommended. To do so, add this line to the
kernel configuration file:
.Bd -literal -offset indent
pseudo-device vinum
.Ed
.Pp
.Ss DEBUG OPTIONS
The current version of
.Nm ,
both the kernel module and the user program
.Xr vinum 8 ,
include significant debugging support. It is not recommended to remove
this support at the moment, but if you do you must remove it from both the
kernel and the user components. To do this, edit the files
.Pa /usr/src/sbin/vinum/Makefile
and
.Pa /usr/src/sys/modules/vinum/Makefile
and edit the CFLAGS variable to remove the -DVINUMDEBUG option. If you have
configured
.Nm
into the kernel, either specify the line
.Bd -literal -offset indent
options VINUMDEBUG
.Ed
.Pp
in the kernel configuration file or remove the -DVINUMDEBUG option from
.Pa /usr/src/sbin/vinum/Makefile
as described above.
.Pp
If the VINUMDEBUG variables do not match,
.Xr vinum 8
will fail with a message
explaining the problem and what to do to correct it.
.Pp
.Nm
was previously available in two versions: a freely available version which did
not contain RAID-5 functionality, and a full version including RAID-5
functionality, which was available only from Cybernet Systems Inc. The present
version of
.Nm
includes the RAID-5 functionality.
.Sh RUNNING VINUM
.Nm
is part of the base
.Fx
system. It does not require installation.
To start it, start the
.Nm
program, which will load the kld if it is not already present.
Before using
.Nm ,
it must be configured. See
.Xr vinum 8
for information on how to create a
.Nm
configuration.
.Pp
Normally, you start a configured version of
.Nm
at boot time. Set the variable
.Ar start_vinum
in
.Pa /etc/rc.conf
to
.Ar YES
to start
.Nm
at boot time.
.Pp
If
.Nm
is loaded as a kld (the recommended way), the
.Nm
.Ar stop
command will unload it. You can also do this with the
.Nm kldunload
command.
.Pp
The kld can only be unloaded when idle, in other words when no volumes are
mounted and no other instances of the
.Nm
program are active. Unloading the kld does not harm the data in the volumes.
.Ss CONFIGURING AND STARTING OBJECTS
Use the
.Xr vinum 8
utility to configure and start
.Nm
objects.
.Sh IOCTL CALLS
.Pa ioctl
calls are intended for the use of the
.Nm
configuration program only. They are described in the header file
.Pa /sys/dev/vinum/vinumio.h
.Ss DISK LABELS
Conventional disk special devices have a
.Em disk label
in the second sector of the device. See
.Xr disklabel 5
for more details. This disk label describes the layout of the partitions within
the device.
.Nm
does not subdivide volumes, so volumes do not contain a physical disk label.
For convenience,
.Nm
implements the ioctl calls DIOCGDINFO (get disk label), DIOCGPART (get partition
information), DIOCWDINFO (write partition information) and DIOCSDINFO (set
partition information). DIOCGDINFO and DIOCGPART refer to an internal
representation of the disk label which is not present on the volume. As a
result, the
.Fl r
option of
.Xr disklabel 8 ,
which reads the
.if t ``raw disk'',
.if n "raw disk",
will fail.
.Pp
In general,
.Xr disklabel 8
serves no useful purpose on a vinum volume. If you run it, it will show you
three partitions, a, b and c, all the same except for the fstype, for example:
.br
.ne 1i
.Bd -literal -offset
3 partitions:
# size offset fstype [fsize bsize bps/cpg]
a: 2048 0 4.2BSD 1024 8192 0 # (Cyl. 0 - 0)
b: 2048 0 swap # (Cyl. 0 - 0)
c: 2048 0 unused 0 0 # (Cyl. 0 - 0)
.Ed
.Pp
.Nm
ignores the DIOCWDINFO and DIOCSDINFO ioctls, since there is nothing to change.
As a result, any attempt to modify the disk label will be silently ignored.
.Sh MAKING FILE SYSTEMS
Since
.Nm
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 partition identifier (a to h).
.Nm
volumes need not conform to this convention, but if they do not,
.Nm newfs
will complain that it cannot determine the partition. To solve this problem,
use the
.Fl v
flag to
.Nm newfs .
For example, if you have a volume
.Pa concat ,
use the following command to create a ufs file system on it:
.Pp
.Bd -literal
# newfs -v /dev/vinum/rconcat
.Ed
.Pp
As with other file systems, you perform the
.Nm newfs
command on the raw device, not the block device.
.Sh OBJECT NAMING
.Nm
assigns default names to plexes and subdisks, although they may be overridden.
We do not recommend overriding the default names. Experience with the
.if t Veritas\(tm
.if n Veritas(tm)
volume manager, which allows arbitary naming of objects, has shown that this
flexibility does not bring a significant advantage, and it can cause confusion.
.sp
Names may contain any non-blank character, but it is recommended to restrict
them to letters, digits and the underscore characters. The names of volumes,
plexes and subdisks may be up to 64 characters long, and the names of drives may
up to 32 characters long. When choosing volume and plex names, bear in mind
that automatically generated plex and subdisk names are longer than the name
from which they are derived.
.Bl -bullet
.It
When
.Xr vinum 8
creates or deletes objects, it creates a directory
.Pa /dev/vinum ,
in which it makes device entries for each volume it finds. It also creates
subdirectories,
.Pa /dev/vinum/plex ,
.Pa /dev/vinum/sd
and
.Pa /dev/vinum/rsd ,
in which it stores device entries for the plexes and subdisks.
.Pa /dev/vinum/sd
contains block device entries, while
.Pa /dev/vinum/rsd
contains character device entries. In addition, it creates two more
directories,
.Pa /dev/vinum/vol
and
.Pa /dev/vinum/drive ,
in which it stores hierarchical information for volumes and drives.
.It
In addition,
.Nm
creates three super-devices,
.Pa /dev/vinum/control ,
.Pa /dev/vinum/Control
and
.Pa /dev/vinum/controld .
.Pa /dev/vinum/control
is used by
.Xr vinum 8
when it has been compiled without the VINUMDEBUG option,
.Pa /dev/vinum/Control
is used by
.Xr vinum 8
when it has been compiled with the VINUMDEBUG option,
and
.Pa /dev/vinum/controld
is used by the
.Nm
daemon. The two control devices for
.Xr vinum 8
are used to synchronize the debug status of kernel and user modules.
.It
Unlike
.Nm UNIX
drives,
.Nm
volumes are not subdivided into partitions, and thus do not contain a disk
label. Unfortunately, this confuses a number of utilities, notably
.Nm newfs ,
which normally tries to interpret the last letter of a
.Nm
volume name as a partition identifier. If you use a volume name which does not
end in the letters
.Ar a
to
.Ar c ,
you must use the
.Fl v
flag to
.Nm newfs
in order to tell it to ignore this convention.
.\"
.It
Plexes do not need to be assigned explicit names. By default, a plex name is
the name of the volume followed by the letters \f(CW.p\fR and the number of the
plex. For example, the plexes of volume
.Ar vol3
are called
.Ar vol3.p0 ,
.Ar vol3.p1
and so on. These names can be overridden, but it is not recommended.
.br
.It
Like plexes, subdisks are assigned names automatically, and explicit naming is
discouraged. A subdisk name is the name of the plex followed by the letters
\f(CW\&.s\fR and a number identifying the subdisk. For example, the subdisks of
plex
.Ar vol3.p0
are called
.Ar vol3.p0.s0 ,
.Ar vol3.p0.s1
and so on.
.br
.It
By contrast,
.Nm drives
must be named. This makes it possible to move a drive to a different location
and still recognize it automatically. Drive names may be up to 32 characters
long.
.El
.Pp
EXAMPLE
.Pp
Assume the
.Nm
objects described in the section CONFIGURATION FILE in
.Xr vinum 8 .
The directory
.Ar /dev/vinum
looks like:
.Bd -literal -offset indent
# ls -lR /dev/vinum/ /dev/rvinum
total 5
brwxr-xr-- 1 root wheel 25, 2 Mar 30 16:08 concat
brwx------ 1 root wheel 25, 0x40000000 Mar 30 16:08 control
brwx------ 1 root wheel 25, 0x40000001 Mar 30 16:08 controld
drwxrwxrwx 2 root wheel 512 Mar 30 16:08 drive
drwxrwxrwx 2 root wheel 512 Mar 30 16:08 plex
drwxrwxrwx 2 root wheel 512 Mar 30 16:08 rvol
drwxrwxrwx 2 root wheel 512 Mar 30 16:08 sd
brwxr-xr-- 1 root wheel 25, 3 Mar 30 16:08 strcon
brwxr-xr-- 1 root wheel 25, 1 Mar 30 16:08 stripe
brwxr-xr-- 1 root wheel 25, 0 Mar 30 16:08 tinyvol
drwxrwxrwx 7 root wheel 512 Mar 30 16:08 vol
brwxr-xr-- 1 root wheel 25, 4 Mar 30 16:08 vol5
/dev/vinum/drive:
total 0
brw-r----- 1 root operator 4, 15 Oct 21 16:51 drive2
brw-r----- 1 root operator 4, 31 Oct 21 16:51 drive4
/dev/vinum/plex:
total 0
brwxr-xr-- 1 root wheel 25, 0x10000002 Mar 30 16:08 concat.p0
brwxr-xr-- 1 root wheel 25, 0x10010002 Mar 30 16:08 concat.p1
brwxr-xr-- 1 root wheel 25, 0x10000003 Mar 30 16:08 strcon.p0
brwxr-xr-- 1 root wheel 25, 0x10010003 Mar 30 16:08 strcon.p1
brwxr-xr-- 1 root wheel 25, 0x10000001 Mar 30 16:08 stripe.p0
brwxr-xr-- 1 root wheel 25, 0x10000000 Mar 30 16:08 tinyvol.p0
brwxr-xr-- 1 root wheel 25, 0x10000004 Mar 30 16:08 vol5.p0
brwxr-xr-- 1 root wheel 25, 0x10010004 Mar 30 16:08 vol5.p1
/dev/vinum/rvol:
total 0
crwxr-xr-- 1 root wheel 91, 2 Mar 30 16:08 concat
crwxr-xr-- 1 root wheel 91, 3 Mar 30 16:08 strcon
crwxr-xr-- 1 root wheel 91, 1 Mar 30 16:08 stripe
crwxr-xr-- 1 root wheel 91, 0 Mar 30 16:08 tinyvol
crwxr-xr-- 1 root wheel 91, 4 Mar 30 16:08 vol5
/dev/vinum/sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20000002 Mar 30 16:08 concat.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100002 Mar 30 16:08 concat.p0.s1
brwxr-xr-- 1 root wheel 25, 0x20010002 Mar 30 16:08 concat.p1.s0
brwxr-xr-- 1 root wheel 25, 0x20000003 Mar 30 16:08 strcon.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100003 Mar 30 16:08 strcon.p0.s1
brwxr-xr-- 1 root wheel 25, 0x20010003 Mar 30 16:08 strcon.p1.s0
brwxr-xr-- 1 root wheel 25, 0x20110003 Mar 30 16:08 strcon.p1.s1
brwxr-xr-- 1 root wheel 25, 0x20000001 Mar 30 16:08 stripe.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100001 Mar 30 16:08 stripe.p0.s1
brwxr-xr-- 1 root wheel 25, 0x20000000 Mar 30 16:08 tinyvol.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100000 Mar 30 16:08 tinyvol.p0.s1
brwxr-xr-- 1 root wheel 25, 0x20000004 Mar 30 16:08 vol5.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100004 Mar 30 16:08 vol5.p0.s1
brwxr-xr-- 1 root wheel 25, 0x20010004 Mar 30 16:08 vol5.p1.s0
brwxr-xr-- 1 root wheel 25, 0x20110004 Mar 30 16:08 vol5.p1.s1
/dev/vinum/vol:
total 5
brwxr-xr-- 1 root wheel 25, 2 Mar 30 16:08 concat
drwxr-xr-x 4 root wheel 512 Mar 30 16:08 concat.plex
brwxr-xr-- 1 root wheel 25, 3 Mar 30 16:08 strcon
drwxr-xr-x 4 root wheel 512 Mar 30 16:08 strcon.plex
brwxr-xr-- 1 root wheel 25, 1 Mar 30 16:08 stripe
drwxr-xr-x 3 root wheel 512 Mar 30 16:08 stripe.plex
brwxr-xr-- 1 root wheel 25, 0 Mar 30 16:08 tinyvol
drwxr-xr-x 3 root wheel 512 Mar 30 16:08 tinyvol.plex
brwxr-xr-- 1 root wheel 25, 4 Mar 30 16:08 vol5
drwxr-xr-x 4 root wheel 512 Mar 30 16:08 vol5.plex
/dev/vinum/vol/concat.plex:
total 2
brwxr-xr-- 1 root wheel 25, 0x10000002 Mar 30 16:08 concat.p0
drwxr-xr-x 2 root wheel 512 Mar 30 16:08 concat.p0.sd
brwxr-xr-- 1 root wheel 25, 0x10010002 Mar 30 16:08 concat.p1
drwxr-xr-x 2 root wheel 512 Mar 30 16:08 concat.p1.sd
/dev/vinum/vol/concat.plex/concat.p0.sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20000002 Mar 30 16:08 concat.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100002 Mar 30 16:08 concat.p0.s1
/dev/vinum/vol/concat.plex/concat.p1.sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20010002 Mar 30 16:08 concat.p1.s0
/dev/vinum/vol/strcon.plex:
total 2
brwxr-xr-- 1 root wheel 25, 0x10000003 Mar 30 16:08 strcon.p0
drwxr-xr-x 2 root wheel 512 Mar 30 16:08 strcon.p0.sd
brwxr-xr-- 1 root wheel 25, 0x10010003 Mar 30 16:08 strcon.p1
drwxr-xr-x 2 root wheel 512 Mar 30 16:08 strcon.p1.sd
/dev/vinum/vol/strcon.plex/strcon.p0.sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20000003 Mar 30 16:08 strcon.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100003 Mar 30 16:08 strcon.p0.s1
/dev/vinum/vol/strcon.plex/strcon.p1.sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20010003 Mar 30 16:08 strcon.p1.s0
brwxr-xr-- 1 root wheel 25, 0x20110003 Mar 30 16:08 strcon.p1.s1
/dev/vinum/vol/stripe.plex:
total 1
brwxr-xr-- 1 root wheel 25, 0x10000001 Mar 30 16:08 stripe.p0
drwxr-xr-x 2 root wheel 512 Mar 30 16:08 stripe.p0.sd
/dev/vinum/vol/stripe.plex/stripe.p0.sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20000001 Mar 30 16:08 stripe.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100001 Mar 30 16:08 stripe.p0.s1
/dev/vinum/vol/tinyvol.plex:
total 1
brwxr-xr-- 1 root wheel 25, 0x10000000 Mar 30 16:08 tinyvol.p0
drwxr-xr-x 2 root wheel 512 Mar 30 16:08 tinyvol.p0.sd
/dev/vinum/vol/tinyvol.plex/tinyvol.p0.sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20000000 Mar 30 16:08 tinyvol.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100000 Mar 30 16:08 tinyvol.p0.s1
/dev/vinum/vol/vol5.plex:
total 2
brwxr-xr-- 1 root wheel 25, 0x10000004 Mar 30 16:08 vol5.p0
drwxr-xr-x 2 root wheel 512 Mar 30 16:08 vol5.p0.sd
brwxr-xr-- 1 root wheel 25, 0x10010004 Mar 30 16:08 vol5.p1
drwxr-xr-x 2 root wheel 512 Mar 30 16:08 vol5.p1.sd
/dev/vinum/vol/vol5.plex/vol5.p0.sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20000004 Mar 30 16:08 vol5.p0.s0
brwxr-xr-- 1 root wheel 25, 0x20100004 Mar 30 16:08 vol5.p0.s1
/dev/vinum/vol/vol5.plex/vol5.p1.sd:
total 0
brwxr-xr-- 1 root wheel 25, 0x20010004 Mar 30 16:08 vol5.p1.s0
brwxr-xr-- 1 root wheel 25, 0x20110004 Mar 30 16:08 vol5.p1.s1
/dev/rvinum:
crwxr-xr-- 1 root wheel 91, 2 Mar 30 16:08 rconcat
crwxr-xr-- 1 root wheel 91, 3 Mar 30 16:08 rstrcon
crwxr-xr-- 1 root wheel 91, 1 Mar 30 16:08 rstripe
crwxr-xr-- 1 root wheel 91, 0 Mar 30 16:08 rtinyvol
crwxr-xr-- 1 root wheel 91, 4 Mar 30 16:08 rvol5
.Ed
.Pp
In the case of unattached plexes and subdisks, the naming is reversed. Subdisks
are named after the disk on which they are located, and plexes are named after
the subdisk.
.\" XXX
.Nm This mapping is still to be determined.
.Ss OBJECT STATES
.Pp
Each
.Nm
object has a \fIstate\fR associated with it.
.Nm
uses this state to determine the handling of the object.
.Pp
.Ss VOLUME STATES
Volumes may have the following states:
.sp
.Bl -hang -width 14n
.It Li down
The volume is completely inaccessible.
.It Li up
The volume is up and at least partially functional. Not all plexes may be
available.
.El
.Ss "PLEX STATES"
Plexes may have the following states:
.sp
.ne 1i
.Bl -hang -width 14n
.It Li referenced
A plex entry which has been referenced as part of a volume, but which is
currently not known.
.It Li faulty
A plex which has gone completely down because of I/O errors.
.It Li down
A plex which has been taken down by the administrator.
.It Li initializing
A plex which is being initialized.
.sp
The remaining states represent plexes which are at least partially up.
.It Li corrupt
A plex entry which is at least partially up. Not all subdisks are available,
and an inconsistency has occurred. If no other plex is uncorrupted, the volume
is no longer consistent.
.It Li degraded
A RAID-5 plex entry which is accessible, but one subdisk is down, requiring
recovery for many I/O requests.
.It Li flaky
A plex which is really up, but which has a reborn subdisk which we don't
completely trust, and which we don't want to read if we can avoid it.
.It Li up
A plex entry which is completely up. All subdisks are up.
.El
.sp 2v
.Ss "SUBDISK STATES"
Subdisks can have the following states:
.sp
.ne 1i
.Bl -hang -width 14n
.It Li empty
A subdisk entry which has been created completely. All fields are correct, and
the disk has been updated, but the on the disk is not valid.
.It Li referenced
A subdisk entry which has been referenced as part of a plex, but which is
currently not known.
.It Li initializing
A subdisk entry which has been created completely and which is currently being
initialized.
.sp
The following states represent invalid data.
.It Li obsolete
A subdisk entry which has been created completely. All fields are correct, the
config on disk has been updated, and the data was valid, but since then the
drive has been taken down, and as a result updates have been missed.
.It Li stale
A subdisk entry which has been created completely. All fields are correct, the
disk has been updated, and the data was valid, but since then the drive has been
crashed and updates have been lost.
.sp
The following states represent valid, inaccessible data.
.It Li crashed
A subdisk entry which has been created completely. All fields are correct, the
disk has been updated, and the data was valid, but since then the drive has gone
down. No attempt has been made to write to the subdisk since the crash, so the
data is valid.
.It Li down
A subdisk entry which was up, which contained valid data, and which was taken
down by the administrator. The data is valid.
.It Li reviving
The subdisk is currently in the process of being revived. We can write but not
read.
.sp
The following states represent accessible subdisks with valid data.
.It Li reborn
A subdisk entry which has been created completely. All fields are correct, the
disk has been updated, and the data was valid, but since then the drive has gone
down and up again. No updates were lost, but it is possible that the subdisk
has been damaged. We won't read from this subdisk if we have a choice. If this
is the only subdisk which covers this address space in the plex, we set its
state to up under these circumstances, so this status implies that there is
another subdisk to fulfil the request.
.It Li up
A subdisk entry which has been created completely. All fields are correct, the
disk has been updated, and the data is valid.
.El
.sp 2v
.Ss "DRIVE STATES"
Drives can have the following states:
.sp
.ne 1i
.Bl -hang -width 14n
.It Li referenced
At least one subdisk refers to the drive, but it is not currently accessible to
the system. No device name is known.
.It Li down
The drive is not accessible.
.It Li up
The drive is up and running.
.El
.sp 2v
.Sh BUGS
.Bl -enum
.It
.Nm
is a new product. Bugs can be expected. The configuration mechanism is not yet
fully functional. If you have difficulties, please look at the section
DEBUGGING PROBLEMS WITH VINUM before reporting problems.
.It
Kernels with the
.Nm
pseudo-device appear to work, but are not supported. If you have trouble with
this configuration, please first replace the kernel with a non-Vinum
kernel and test with the kld module.
.It
Detection of differences between the version of the kernel and the kld is not
yet implemented.
.It
The RAID-5 functionality is new in
.Fx 3.3 .
Some problems have been
reported with
.Nm
in combination with soft updates, but these are not reproducible on all
systems. If you are planning to use
.Nm
in a production environment, please test carefully.
.El
.Sh DEBUGGING PROBLEMS WITH VINUM
.Pp
Solving problems with
.Nm
can be a difficult affair. This section suggests some approaches.
.Ss Configuration problems
.Pp
It is relatively easy (too easy) to run into problems with the
.Nm
configuration. If you do, the first thing you should do is stop configuration
updates:
.if t .ps -3
.if t .vs -3
.Bd -literal
# vinum setdaemon 4
.Ed
.if t .vs
.if t .ps
.Pp
This will stop updates and any further corruption of the on-disk configuration.
.Pp
Next, look at the on-disk configuration, using a Bourne-style shell:
.if t .ps -3
.if t .vs -3
.Bd -literal
# rm -f log
# for i in /dev/da0s1h /dev/da1s1h /dev/da2s1h /dev/da3s1h; do
(dd if=$i skip=8 count=6|tr -d '\e000-\e011\e200-\e377'; echo) >> log
done
.Ed
.if t .vs
.if t .ps
.Pp
The names of the devices are the names of all
.Nm
slices. The file
.Pa log
should then contain something like this:
.if t .ps -3
.if t .vs -3
.Bd -literal
IN VINOpanic.lemis.comdrive1}6E7~^K6T^Yfoovolume obj state up
volume src state up
volume raid state down
volume r state down
volume foo state up
plex name obj.p0 state corrupt org concat vol obj
plex name obj.p1 state corrupt org striped 128b vol obj
plex name src.p0 state corrupt org striped 128b vol src
plex name src.p1 state up org concat vol src
plex name raid.p0 state faulty org disorg vol raid
plex name r.p0 state faulty org disorg vol r
plex name foo.p0 state up org concat vol foo
plex name foo.p1 state faulty org concat vol foo
sd name obj.p0.s0 drive drive2 plex obj.p0 state reborn len 409600b driveoffset 265b plexoffset 0b
sd name obj.p0.s1 drive drive4 plex obj.p0 state up len 409600b driveoffset 265b plexoffset 409600b
sd name obj.p1.s0 drive drive1 plex obj.p1 state up len 204800b driveoffset 265b plexoffset 0b
sd name obj.p1.s1 drive drive2 plex obj.p1 state reborn len 204800b driveoffset 409865b plexoffset 128b
sd name obj.p1.s2 drive drive3 plex obj.p1 state up len 204800b driveoffset 265b plexoffset 256b
sd name obj.p1.s3 drive drive4 plex obj.p1 state up len 204800b driveoffset 409865b plexoffset 384b
.Ed
.if t .vs
.if t .ps
.Pp
The first line contains the
.Nm
label and must start with the text
.Li IN VINO .
It also contains the name of the system. The exact definition is contained in
.Pa /usr/src/sys/dev/vinum/vinumvar.h .
The saved configuration starts in the middle of the line with the text
.Li volume obj state up
and starts in sector 9 of the disk.
The rest of the output shows the remainder of the on-disk configuration. It
may be necessary to increase the
.Ar count
argument of
.Cm dd
in order to see the complete configuration.
.Pp
The configuration on all disks should be the same. If this is not the case,
please report the problem with the exact contents of the file
.Pa log .
There is probably little that can be done to recover the on-disk configuration,
but if you keep a copy of the files used to create the objects, you should be
able to re-create them. The
.Cm create
command does not change the subdisk data, so this will not cause data
corruption. You may need to use the
.Cm resetconfig
command if you have this kind of trouble.
.Ss Kernel Panics
.Pp
In order to analyse a panic which you suspect comes from
.Nm
you will need to build a debug kernel. See the online handbook at
.Pa /usr/share/doc/handbook/kerneldebug.html
(if installed) or
.Pa http://www.FreeBSD.org/handbook/kerneldebug.html
for more details of how to do this.
.Pp
Perform the following steps to analyse a
.Nm
problem:
.Bl -enum
.It
Copy the files
.Pa /usr/src/sys/modules/vinum/.gdbinit.crash ,
.Pa /usr/src/sys/modules/vinum/.gdbinit.kernel ,
.Pa /usr/src/sys/modules/vinum/.gdbinit.serial ,
.Pa /usr/src/sys/modules/vinum/.gdbinit.vinum
and
.Pa /usr/src/sys/modules/vinum/.gdbinit.vinum.paths
to the directory in which you will be performing the analysis, typically
.Pa /var/crash .
.It
Make sure that you build the
.Nm
module with debugging information. The standard
.Pa Makefile
builds a module with debugging symbols by default. If the version of
.Nm
in
.Pa /modules
does not contain symbols, you will not get an error message, but the stack trace
will not show the symbols. Check the module before starting
.Nm gdb :
.Bd -literal
$ file /modules/vinum.ko
/modules/vinum.ko: ELF 32-bit LSB shared object, Intel 80386,
version 1 (FreeBSD), not stripped
.Ed
.Pp
If the output shows that
.Pa /modules/vinum.ko
is stripped, you will have to find a version which is not. Usually this will be
either in
.Pa /usr/obj/sys/modules/vinum/vinum.ko
(if you have built
.Nm
with a
.Ar make world )
or
.Pa /usr/src/sys/modules/vinum/vinum.ko
(if you have built
.Nm
in this directory). Modify the file
.Pa .gdbinit.vinum.paths
accordingly.
.It
Either take a dump or use remote serial
.Cm gdb
to analyse the problem. To analyse a dump, say
.Pa /var/crash/vmcore.5 ,
link
.Pa /var/crash/.gdbinit.crash
to
.Pa /var/crash/.gdbinit
and enter:
.Bd -literal
# cd /var/crash
# gdb -k kernel.debug vmcore.5
.Ed
.Pp
This example assumes that you have installed the correct debug kernel at
.Pa /var/crash/kernel.debug .
If not, substitute the correct name of the debug kernel.
.Pp
To perform remote serial debugging,
link
.Pa /var/crash/.gdbinit.serial
to
.Pa /var/crash/.gdbinit
and enter
.Bd -literal
# cd /var/crash
# gdb -k kernel.debug
.Ed
.Pp
In this case, the
.Pa .gdbinit
file performs the functions necessary to establish connection. The remote
machine must already be in debug mode: enter the kernel debugger and select
.Nm gdb .
The serial
.Pa .gdbinit
file expects the serial connection to run at 38400 bits per second; if you run
at a different speed, edit the file accordingly (look for the
.Ar remotebaud
specification).
.Pp
The following example shows a remote debugging session using the
.Ar debug
command of
.Xr vinum 8 :
.if t .ps -3
.if t .vs -3
.Bd -literal
GDB 4.16 (i386-unknown-freebsd), Copyright 1996 Free Software Foundation, Inc.
Debugger (msg=0xf1093174 "vinum debug") at ../../i386/i386/db_interface.c:318
318 in_Debugger = 0;
#1 0xf108d9bc in vinumioctl (dev=0x40001900, cmd=0xc008464b, data=0xf6dedee0 "",
flag=0x3, p=0xf68b7940) at
/usr/src/sys/modules/Vinum/../../dev/Vinum/vinumioctl.c:102
102 Debugger ("vinum debug");
(kgdb) bt
#0 Debugger (msg=0xf0f661ac "vinum debug") at ../../i386/i386/db_interface.c:318
#1 0xf0f60a7c in vinumioctl (dev=0x40001900, cmd=0xc008464b, data=0xf6923ed0 "",
flag=0x3, p=0xf688e6c0) at
/usr/src/sys/modules/vinum/../../dev/vinum/vinumioctl.c:109
#2 0xf01833b7 in spec_ioctl (ap=0xf6923e0c) at ../../miscfs/specfs/spec_vnops.c:424
#3 0xf0182cc9 in spec_vnoperate (ap=0xf6923e0c) at ../../miscfs/specfs/spec_vnops.c:129
#4 0xf01eb3c1 in ufs_vnoperatespec (ap=0xf6923e0c) at ../../ufs/ufs/ufs_vnops.c:2312
#5 0xf017dbb1 in vn_ioctl (fp=0xf1007ec0, com=0xc008464b, data=0xf6923ed0 "",
p=0xf688e6c0) at vnode_if.h:395
#6 0xf015dce0 in ioctl (p=0xf688e6c0, uap=0xf6923f84) at ../../kern/sys_generic.c:473
#7 0xf0214c0b in syscall (frame={tf_es = 0x27, tf_ds = 0x27, tf_edi = 0xefbfcff8,
tf_esi = 0x1, tf_ebp = 0xefbfcf90, tf_isp = 0xf6923fd4, tf_ebx = 0x2,
tf_edx = 0x804b614, tf_ecx = 0x8085d10, tf_eax = 0x36, tf_trapno = 0x7,
tf_err = 0x2, tf_eip = 0x8060a34, tf_cs = 0x1f, tf_eflags = 0x286,
tf_esp = 0xefbfcf78, tf_ss = 0x27}) at ../../i386/i386/trap.c:1100
#8 0xf020a1fc in Xint0x80_syscall ()
#9 0x804832d in ?? ()
#10 0x80482ad in ?? ()
#11 0x80480e9 in ?? ()
.Ed
.if t .vs
.if t .ps
.Pp
When entering from the debugger, it's important that the source of frame 1
(listed by the
.Pa .gdbinit
file at the top of the example) contains the text
.if t .ps -3
.if t .vs -3
.Bd -literal
Debugger ("vinum debug");
.Ed
.if t .vs
.if t .ps
.Pp
This is an indication that the address specifications are correct. If you get
some other output, your symbols and the kernel module are out of sync, and the
trace will be meaningless.
.El
.Pp
For an initial investigation, the most important information is the output of
the
.Nm bt
(backtrace) command above.
.Ss Reporting problems with Vinum
.Pp
If you find any bugs in
.Nm ,
please report them to Greg Lehey <grog@lemis.com>. Supply the following
information:
.Pp
.Bl -bullet
.It
The output of the
.Nm
.Cm list
command.
.It
Any messages printed in
.Pa /var/log/messages .
All such messages will be identified by the text
.Nm
at the beginning.
.It
If you have a panic, a stack trace as described above.
.El
.Sh AUTHORS
.An Greg Lehey Aq grog@lemis.com .
.Sh HISTORY
.Nm
first appeared in
.Fx 3.0 .
The RAID-5 component of
.Nm
was developed by Cybernet Inc.
.Pa www.cybernet.com
for its NetMAX product.
.Sh SEE ALSO
.Xr vinum 8 ,
.Xr disklabel 5 ,
.Xr disklabel 8 ,
.Xr newfs 8