On Thu, 2002-08-29 at 18:23, Greg Freemyer wrote:
>
> Nathan,
>
> I have made a second attempt at the script.
>
> I think it addresses your concerns.
>
> Someone still needs to do the TODO items, but they don't come up in my
> environment because $SCRATCH_DEV is a LV for me.
>
> Steve Lord has run my previous version to some extent, but I don't know if he
> did it via check, or he just ran the script.
>
I ran it by hand, I think check needs to know about the specific test
numbers, it also wants a good output file to compare against.
One thing I did discover was the partition type needs to be set
correctly. I presume that was what the dd from /dev/zero was for,
I only had a free partition on a disk, not a free volume.
I have subsequently been chasing oopses from running freeze/thaw
on a filesystem under heavy load, but that is a different problem
than you saw I think.
Steve
> Greg Freemyer
> Internet Engineer
> Deployment and Integration Specialist
> Compaq ASE - Tru64 v4, v5
> Compaq Master ASE - SAN Architect
> The Norcross Group
> www.NorcrossGroup.com
>
> ======== 068.out
> QA output created by 068
> SUCCESS, COMPLETED ALL ITERATIONS WITH NO TIME OUTS!!!!!!!!!!!!
> Cleanup beginning
> ======== 068
> #! /bin/sh
> # XFS QA Test No. 068
> # $Id: 1.1 $
> #
> # Test LVM snapshot creation
> #
> # The timing and placement of kills and waits is particularily sensitive.
> # Don't change them unless you want to spend some time getting it right
> again.
> #
> #-----------------------------------------------------------------------
> # Copyright (c) 2000, 2002 Silicon Graphics, Inc. All Rights Reserved.
> #
> # This program is free software; you can redistribute it and/or modify it
> # under the terms of version 2 of the GNU General Public License as
> # published by the Free Software Foundation.
> #
> # This program is distributed in the hope that it would be useful, but
> # WITHOUT ANY WARRANTY; without even the implied warranty of
> # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
> #
> # Further, this software is distributed without any warranty that it is
> # free of the rightful claim of any third person regarding infringement
> # or the like. Any license provided herein, whether implied or
> # otherwise, applies only to this software file. Patent licenses, if
> # any, provided herein do not apply to combinations of this program with
> # other software, or any other product whatsoever.
> #
> # You should have received a copy of the GNU General Public License along
> # with this program; if not, write the Free Software Foundation, Inc., 59
> # Temple Place - Suite 330, Boston MA 02111-1307, USA.
> #
> # Contact information: Silicon Graphics, Inc., 1600 Amphitheatre Pkwy,
> # Mountain View, CA 94043, or:
> #
> # http://www.sgi.com
> #
> # For further information regarding this notice, see:
> #
> # http://oss.sgi.com/projects/GenInfo/SGIGPLNoticeExplan/
> #-----------------------------------------------------------------------
> #
> # creator
> owner=freemyer@xxxxxxxxxxxxxxxxx
>
> seq=`basename $0`
> echo "QA output created by $seq"
>
> here=`pwd`
> tmp=/tmp/$$
> status=1 # failure is the default!
>
> GENERATE_IO_LOAD=TRUE # If "FALSE", the dd load loop is skipped
> DELAY_BETWEEN_ITERATIONS=10
> ITERATIONS=20
> VG=/dev/VGscratch
> #SCRATCH_DEV=/dev/xxxx # Only needed if running by hand, ie. check
> sets these
> #SCRATCH_MNT=/scratch # Only needed if running by hand, ie. check
> sets these
> SCRATCH_SNAP_MNT=$tmp.scratch_snap
>
>
> _cleanup()
> {
>
> echo Cleanup beginning
>
> rm $tmp.running > /dev/null 2>&1
> xfs_freeze -u $SCRATCH_MNT
>
> sleep 10 # Give the dd loop time to finish
>
> # Comment out unless needed. If needed, wrap with logic to ensure
> the FS is mounted
> #Kill off any other possible stray stragglers that may be out there.
> There should not be any.
> # fuser -k -m $SCRATCH_SNAP_MNT/dummy >/dev/null 2>&1
> # fuser -k -m $SCRATCH_MNT/dummy >/dev/null 2>&1
>
> wait
>
> umount $SCRATCH_SNAP_MNT > /dev/null 2>&1
> rmdir $SCRATCH_SNAP_MNT > /dev/null 2>&1
> umount $SCRATCH_MNT > /dev/null 2>&1
>
> lvremove -f $VG/scratch_snap > /dev/null 2>&1
> lvremove -f $VG/scratch > /dev/null 2>&1
>
> #TODO vgremove $VG
>
> rm -f $tmp.* # if we ever use tmp files
> trap 0 1 2 3 15
> exit $status
> }
>
> trap "_cleanup" 0 1 2 3 15
>
>
> # get standard environment, filters and checks
> . ./common.rc
> . ./common.filter
>
> if [ -e $SCRATCH_SNAP_MNT ]; then rm -rf $SCRATCH_SNAP_MNT; fi
>
> mkdir $SCRATCH_SNAP_MNT
>
> #Verify we have the lvm user tools
> [ -x /sbin/lvcreate ] || _notrun "LVM lvcreate utility is not installed in
> /sbin"
> [ -x /sbin/lvremove ] || _notrun "LVM lvremove utility is not installed in
> /sbin"
>
> # if the above fails for LVM 2, the below may be useful notes
> # try "lvm vgdisplay --version" and look for a > 1.9x version number, eg,
> #
> # LVM version: 1.95.10-cvs (2002-05-31)
> # Library version: 0.96.03-ioctl-cvs (2002-06-27)
> # Driver version: 1.0.3
>
>
> #Verify we have the a lvm enabled kernel
>
> LVM=false
>
> #Check if LVM 1 is in the kernel
> if grep lvm /proc/devices > /dev/null 2>&1; then LVM=true; fi
>
> #Check if LVM 2 is in the kernel
> if grep device-mapper /proc/devices > /dev/null 2>&1; then LVM=true; fi
>
> #Check if EVMS is in the kernel
> # TODO # I don't know how to do this one.
>
> if [ $LVM = false ]; then _notrun "This test requires the kernel have LVM or
> EVMS present. (The EVMS test is still TBD)."; fi
>
>
> # real QA test starts here
>
> # Create a PV set from the scratch partition
> # TODO # (I don't know if this is needed.and it is dangerous because it
> intentionally deletes the partition table!!!
> # TODO # dd if=/dev/zero of=$SCRATCH_DEV bs=512 count=1
> # TODO # pvcreate $SCRATCH_DEV
>
> # Create a VG from the PV
> # TODO # vgcreate $VG $SCRATCH_DEV
>
> # Create a LV in the VG to snapshot
> # TODO # lvcreate -L 2G -n scratch $VG
>
> # Mount the LV
> mkdir $SCRATCH_MNT > /dev/null 2&>1
>
> mount $VG/scratch $SCRATCH_MNT
>
> if [ $GENERATE_IO_LOAD != FALSE ];
> then
> # Create a large 64 Meg zero filled file on the LV
> dd if=/dev/zero of=$SCRATCH_MNT/dummy bs=64k count=1000 > /dev/null
> 2>&1
>
> #setup an infinite loop to copy the large file, thus generating heavy
> i/o
>
> touch $tmp.running
>
> while [ -f $tmp.running ]
> do
> dd if=$SCRATCH_MNT/dummy of=$SCRATCH_MNT/junk bs=64k > /dev/null
> 2>&1
> rm $SCRATCH_MNT/junk # This forces metadata updates the
> next time around
> sync
> done &
> fi
>
> ii=1
>
> while [ $ii -le $ITERATIONS ]
> do
>
> # echo $ii Usefull if your are running interactive, but not from
> the xfs test scripts
>
> #if the VFS lock patch is present, the calls to xfs_freeze are
> redundant, but should cause no problems
> # OPTIONAL
> xfs_freeze -f $SCRATCH_MNT
> if [ $? != 0 ] ; then
> echo xfs_freeze -f $SCRATCH_MNT failed
> fi
> (
> lvcreate --snapshot --size 1G --name scratch_snap $VG/scratch
> > /dev/null 2>&1
> ret=$?
> if [ $ret != 0 ] ; then
> echo snapshot creation for $SCRATCH_MNT failed with
> return code $ret
> fi
> ) &
> SNAPSHOT_shell_pid=$!
>
> #if the Snapshot has not completed in ten minutes, kill it
> (
> # I have NOT figured out how to kill the sleep 600
> before it exits naturally.
> # This does not cause a problem, but it clutters the
> ps table.
> sleep 600
> # The kill $TIMEOUT_shell_pid keeps the below from
> occuring
> echo Snapshot Lockup Occured on loop $ii
> xfs_freeze -u $SCRATCH_MNT
> kill $$
> ) &
> TIMEOUT_shell_pid=$!
>
> wait $SNAPSHOT_shell_pid
>
> exec 2> /dev/null # Send the shells stderr to /dev/null
> kill $TIMEOUT_shell_pid #Cancel the timeout
> wait $TIMEOUT_shell_pid # This causes consistent shell
> notification for some unknow reason
> exec 2>&1 # Put it back to the same as stdout
>
> #if the VFS lock patch is present, the calls to xfs_freeze are
> redundant, but should cause no problems
> # OPTIONAL
> xfs_freeze -u $SCRATCH_MNT
> if [ $? != 0 ] ; then
> echo xfs_freeze -u $SCRATCH_MNT failed
> fi
> # MANDANTORY (end)
>
> mount -t xfs -o ro,nouuid $VG/scratch_snap $SCRATCH_SNAP_MNT
> if [ $? != 0 ] ; then
> echo mount for $SCRATCH_SNAP_MNT failed
> fi
> umount $SCRATCH_SNAP_MNT
> if [ $? != 0 ] ; then
> echo umount for $SCRATCH_SNAP_MNT failed
> fi
> lvremove -f $VG/scratch_snap > /dev/null 2>&1
> if [ $? != 0 ] ; then
> echo lvremove for $VG/scratch_snap failed
> fi
>
> ii=`expr $ii + 1`
> sleep $DELAY_BETWEEN_ITERATIONS # The VG seems to need time to
> stabalize between snapshots
> # With LVM 1.0.3 and XFS 1.1, I have tried this at
> 3600 seconds and still had failures
>
> done
>
> # success, all done
> echo SUCCESS, COMPLETED ALL ITERATIONS WITH NO TIME OUTS!!!!!!!!!!!!
> status=0
> _cleanup
> exit 1 # _cleanup should exit, so we should never get here.
>
|