>> Greg - Ethan's is more correct, don't worry about consistency. Either
>> is fine.
>> A couple other suggestions: (did you know what you were signing up for
>> here?) :)
>> Rather than using SCRATCH_DEV for the LVM tests, we probably would need
>> a new SCRATCH_LVM_DEV since they need to be different partition types.
>> I doubt that we want to try to change partition types on the fly. :)
>> Also, if you want to use something other than "dd" to generate fs
>> activity, you could use something like fsstress, which is part of the
>> xfstest suite.
>> -Eric
Nathan,
I hope the below is one that is finally ready to go into cvs.
I have added the ability to set a env. variable to select either fsstress or dd
to create the io load. I have just run a quick test with both enabled.
I have also changed all references from SCRATCH_DEV to SCRATCH_LVM_DEV. These
are still all commented out, so when you try it in your test machines you will
have to figure out the details on those.
I have the env. variables in quotes now per Ethans's request, and I am now
using a temp directory /tmp/xfstest.XXXXXX where XXXXXX is random and unique.
The 068.out has not changed, but it is below regardless.
Thanks
Greg Freemyer
========= 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 $seq"
here=`pwd`
tmp=`mktemp -d /tmp/xfstest.XXXXXX` || exit 1
status=1 # failure is the default!
# Either "dd" or "fsstress" can be used to generate io load. Setting both to
"FALSE" tests with no io load.
GENERATE_IO_LOAD_VIA_DD=TRUE
GENERATE_IO_LOAD_VIA_FSSTRESS=TRUE
DELAY_BETWEEN_ITERATIONS=10
#ITERATIONS=10000
ITERATIONS=2
VG=/dev/VGscratch
#SCRATCH_LVM_DEV=/dev/xxxx # Only needed if running by hand, ie.
check sets these
#SCRATCH_MNT=/scratch # Only needed if running by hand, ie. check
sets these
SCRATCH_SNAP_MNT="$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 -rf "$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_LVM_DEV" bs=512 count=1
# TODO # pvcreate "$SCRATCH_LVM_DEV"
# Create a VG from the PV
# TODO # vgcreate "$VG" "$SCRATCH_LVM_DEV"
# Create a LV in the VG to snapshot
# TODO # lvcreate -L 2G -n scratch "$VG"
# Mount the LV
mkdir "$SCRATCH_MNT" > /dev/null 2&>1
mount "$VG/scratch" "$SCRATCH_MNT"
touch "$tmp/running"
if [ "$GENERATE_IO_LOAD_VIA_DD" != FALSE ];
then
# Create a large 64 Meg zero filled file on the LV
dd if=/dev/zero of="$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 "$SCRATCH_MNT/junk" # This forces metadata updates the next
time around
sync
done
fi &
if [ "$GENERATE_IO_LOAD_VIA_FSSTRESS" != FALSE ];
then
STRESS_DIR="$SCRATCH_MNT/fsstress_test_dir"
mkdir "$STRESS_DIR"
while [ -f "$tmp/running" ]
do
# -n 10 makes this take about 10 seconds,
# this allows the loop to end shortly after $tmp/running is
deleted
src/fsstress -d "$STRESS_DIR" -n 10 > /dev/null 2>&1
sync
done
rm -rf "$STRESS_DIR"
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.
TruStor
|