Merge branch 'linux-have-mntent' into refs/top-bases/linux-zfs-branch

This commit is contained in:
Brian Behlendorf 2010-05-21 15:30:47 -07:00
commit 733007a14c
5 changed files with 135 additions and 83 deletions

View File

@ -1,3 +1,47 @@
2010-05-21 Brian Behlendorf <behlendorf1@llnl.gov>
* : Tag zfs-0.4.9 - Use 'git log --no-merges' for full change log.
* : Build system improvements:
- Added support for the 'make -s' silent build option.
- Allow zfs_config.h to be included by dependent packages.
- Minor spec file updates.
- Minor build system message updates.
* : Topic branch cleanup. Several old branches were removed and
numerous hunks which were accidentaly commited to incorrect topic
branches in the past were relocated to the correct topic branch.
* *.c, *.h, *.sh, AUTHORS, COPYING, COPYRIGHT, DISCLAIMER, README:
Added standard header to source files which includes the copyright,
license, and author information. Additionally, updated the listed
top level files to the latest versions.
* cmd/zpool/zpool_vdev.c: Check all partitions with check_file()
even when no libblkid is found. This oversight would result in
ZFS not detecting existing filesystems on partitions.
* module/zfs/fm.c, module/zfs/zfs_fm, libzfs/libzfs_pool.c:
Added zevents which are similar to Solaris FMA support. The
existing FMA and sysevent call points in ZFS were unified in
to a single event type and used to create a user space visible
event notification system under Linux. The new 'zpool events'
command can be used to show all recent ZFS events.
* module/zfs/spa.c, module/zfs/zil.c: Suppress large memory
allocation warnings for two particular kmem_alloc()'s. For
now we can live with them as is but long term a way should be
found to perform small allocations or use the vmem based slab.
* module/zfs/zvol.c: Recreate volume and snapshot /dev links during
module load. Links in /dev for volumes/snapshots were only being
created at volume/snapshot creation time. Those links are now also
created dynamically at module load time based on the spa config.
* module/zfs/zvol.c, module/zfs/include/sys/blkdev.h: The handler
zvol_request() should use the unlocked version of blk_end_request()
to avoid a deadlock in the ZVOL.
2010-03-11 Brian Behlendorf <behlendorf1@llnl.gov> 2010-03-11 Brian Behlendorf <behlendorf1@llnl.gov>
* : Tag zfs-0.4.8 - Use 'git log --no-merges' for full change log. * : Tag zfs-0.4.8 - Use 'git log --no-merges' for full change log.

2
META
View File

@ -1,7 +1,7 @@
Meta: 1 Meta: 1
Name: zfs Name: zfs
Branch: 1.0 Branch: 1.0
Version: 0.4.8 Version: 0.4.9
Release: 1 Release: 1
Release-Tags: relext Release-Tags: relext
License: CDDL License: CDDL

115
README
View File

@ -1,40 +1,95 @@
============================ ZFS KERNEL BUILD ============================ ============================ ZFS QUICK START ============================
1) Build the SPL (Solaris Porting Layer) module which is designed to 1) Build the SPL (Solaris Porting Layer) and install it. This package
provide many Solaris APIs in the Linux kernel which are needed provides several Solaris APIs used by ZFS and is a required dependency.
by ZFS. To build the SPL: Before building ZFS build this package and install the resulting rpms.
tar -xzf spl-x.y.z.tgz > tar -xzf spl-x.y.z.tgz
cd spl-x.y.z > cd spl-x.y.z
./configure --with-linux=<kernel src> > ./configure --with-linux=<kernel src>
make > make
make check <as root> > make rpm
2) Build ZFS, this port is based on build specified by the ZFS.RELEASE > sudo rpm -Uvh *.<arch>.rpm
file. You will need to have both the kernel and SPL source available. Preparing... ########################################### [100%]
To build ZFS for use as a Linux kernel module. 1:spl ########################################### [ 33%]
2:spl-modules-devel ########################################### [ 67%]
3:spl-modules ########################################### [100%]
tar -xzf zfs-x.y.z.tgz
cd zfs-x.y.z
./configure --with-linux=<kernel src> \
--with-spl=<spl src>
make
make check <as root>
============================ ZPIOS TEST SUITE ============================ 2) Build ZFS and install it. This package provides the native port of
ZFS for Linux including all kernel modules and command line utilities.
Note it is important that you have installed spl-module-devel package
from step 1) before attempting to build ZFS.
3) Provided is an in-kernel test application called zpios which can be > tar -xzf zfs-x.y.z.tgz
used to simulate a parallel IO load. It may be used as a stress > cd zfs-x.y.z
or performance test for your configuration. To simplify testing > ./configure --with-linux=<kernel src>
scripts provided in the scripts/ directory which provide a few > make
pre-built zpool configurations and zpios test cases. By default > make rpm
'make check' as root will run a simple test against several small
loopback devices created in /tmp/. > sudo rpm -Uvh *.<arch>.rpm
Preparing... ########################################### [100%]
1:zfs ########################################### [ 20%]
2:zfs-test ########################################### [ 40%]
3:zfs-modules-devel ########################################### [ 60%]
4:zfs-modules ########################################### [ 80%]
3) Enjoy ZFS on Linux! Currently only the ZVOL is fully functional
but work on the ZPL is underway. Why is just having the ZVOL still
useful you ask? Well here's an example of what you can do today using
just the ZVOL:
> # Create the 'tank' zpool containing a raidz vdev spread over 4 devices.
> zpool create tank raidz /dev/sdb /dev/sdc /dev/sdd /dev/sde
> zpool list
NAME SIZE USED AVAIL CAP HEALTH ALTROOT
tank 1.81T 132K 1.81T 0% ONLINE -
> # Create a 100G block device named 'fish' in the 'tank' zpool.
> zfs create -V 100G tank/fish
> zfs list
NAME USED AVAIL REFER MOUNTPOINT
tank 100G 1.24T 26.9K /tank
tank/fish 100G 1.33T 23.9K -
> # Partition 'tank/fish' as if it were a normal block device.
> sfdisk /dev/tank/fish << EOF
0,
EOF
> sfdisk -l /dev/tank/fish
Disk /dev/tank/fish: 208050 cylinders, 16 heads, 63 sectors/track
Units = cylinders of 516096 bytes, blocks of 1024 bytes, counting from 0
Device Boot Start End #cyls #blocks Id System
/dev/tank/fish1 0+ 208049 208050- 104857199+ 83 Linux
/dev/tank/fish2 0 - 0 0 0 Empty
/dev/tank/fish3 0 - 0 0 0 Empty
/dev/tank/fish4 0 - 0 0 0 Empty
> # Format the new /dev/tank/fish1 partition with ext2 and mount it.
> mkfs.ext2 -q /dev/tank/fish1
> mkdir -p /mnt/tank/fish1
> mount /dev/tank/fish1 /mnt/tank/fish1
> ls /mnt/tank/fish1
lost+found
> # Take a snapshot of the pristine ext2 filesystem and mount it read-only.
> zfs snapshot tank/fish@pristine
> mkdir /mnt/tank/fish@pristine1
> mount /dev/tank/fish@pristine1 /mnt/tank/fish@pristine1
> ls /mnt/tank/fish\@pristine1
lost+found
> # Changes made to tank/fish1 do not appear in tank/fish@pristine1
> touch /mnt/tank/fish1/foo
> ls /mnt/tank/fish1/
foo lost+found
> ls /mnt/tank/fish\@pristine1
lost+found
cd scripts
./zfs.sh # Load the ZFS/SPL modules
./zpios.sh -c lo-raid0.sh -t tiny -v # Tiny zpios loopback test
./zfs.sh -u # Unload the ZFS/SPL modules
Enjoy, Enjoy,
Brian Behlendorf <behlendorf1@llnl.gov> Brian Behlendorf <behlendorf1@llnl.gov>

51
TODO
View File

@ -1,51 +0,0 @@
SUMMARY OF MAJOR KNOWN PROBLEMS IN v0.4.6 (Development Release)
* Fault Management (FM) and sysevent support / analog.
  bugzilla 14866, 15645
This is probably the biggest remaining chunk of work.  Linux has no
direct equivalent of the Solaris Fault Management Architecture (FMA)
and we need one.  All fault information is currently ignored and no
disk errors are even logged.  We need to settle on a design for this
but minimally it needs to log the events to syslog.
* Implement the ZVOL.
bugzilla xxxxx
This should be pretty staight forward now that the DMU is fully
implemented and solid. It just needs to be done.
* Implement the ZPL.
bugzilla xxxxx
Getting basic ZPL support should be pretty straight forward. Moving
beyond that to fully integrate with the VFS for things like mmap and
file locking will be trickier.
* Integrate the ZFS-FUSE port in to this code base.
bugzilla xxxxx
Merging the zfs-fuse code base in with this project would be nice from a
code maintence standpoint. This code base is quite a bit newer than
zfs-fuse and it already provides a libzpool library for zfs-fuse to link
against. This should be a pretty straight forward addition.
* Emulate kthreads with pthreads in userspace.
  bugzilla xxxxx
There is a patch available for this but each time I've integrated it
I've observed SIGSEGVs in ztest. Once this patch is in place ztest
can be used to use the kthread API which brings us one step closer
to being able to run it in the kernel as an additional sanity check.
* DMU Performance
  bugzilla 13566
While performance is currently not bad it is not where it needs to be
for production use. The latest test results which can be found in the
docs directly show that on hardware which is capable of 8GB/s we only
see a few GB/s when running through the DMU. To address this we need
to finish getting the code working with the kernel lock profiler and
look for some hot locks. Additionally, it would be interesting to run
the same tests on Solaris (once we have a ZVOL/ZPL) and compare the
performance. It's not at all clear to me Solaris currently does better.

View File

@ -94,7 +94,11 @@
%endif %endif
%define kpkg kernel %define kpkg kernel
%define kdevpkg kernel-devel %define kdevpkg kernel-devel
%if %{defined el6}
%define kverpkg %(echo %{kver} | %{__sed} -e 's/.%{_target_cpu}//g')
%else
%define kverpkg %{kver} %define kverpkg %{kver}
%endif
%define koppkg = %define koppkg =
%if %{undefined kdir} %if %{undefined kdir}
%define kdir %{_usrsrc}/kernels/%{kver}-%{_target_cpu} %define kdir %{_usrsrc}/kernels/%{kver}-%{_target_cpu}