[prev in list] [next in list] [prev in thread] [next in thread]
List: linux-lvm
Subject: [linux-lvm] Draft Snapshot Regression Test
From: Greg Freemyer <freemyer () NorcrossGroup ! com>
Date: 2002-08-28 23:28:13
[Download RAW message or body]
All,
I just sent the below snapshot regression test to the XFS mailing list
It is designed to run from the XFS test harness, but that it should be easy to \
convert to be simply run by hand.
I have only tried it on a LVM 1.0.3 / xfs 1.1 / 2.4.19pre4-aa kernel and it fails \
within 10 iterations every time.
I have seen a couple of successes, but only when I set the iteration count very low, \
like 2 or 3.
On one occasion, the lvcreate locked up so tight, I had to reboot my server to get \
that FS unlocked. (i.e. cycle power because the shutdown command would not run.)
There are a few environmental variables at the beginning you can set to control a \
couple of things.
The script only uses the $SCRATCH_DEV and $SCRATCH_MNT variables from the xfs test \
harness, so you could simply uncomment the currently commented out assignments. The \
lines with _notrun in them can probably be deleted, unless you have a similar test \
harness.
I have not tested any of the TODO lines, but you may or may not need them in your \
test environments.
I copied the license stuff from another SGI test script, but I changed it to my \
company instead of SGI.
Feel free to make changes to this as required.
The script is below my signature.
BTW: This is mostly lvm specific, not xfs specific, so you can likely extend it to \
other filesystem types fairly easily.
Greg Freemyer
Internet Engineer
Deployment and Integration Specialist
Compaq ASE - Tru64 v4, v5
Compaq Master ASE - SAN Architect
The Norcross Group
www.NorcrossGroup.com
=====================================
#! /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 The Norcross Group, 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@NorcrossGroup.com
seq=`basename $0`
echo "QA output created by $seq"
DELAY_BETWEEN_ITERATIONS=20
ITERATIONS=30
VG=/dev/VGscratch
#SCRATCH_DEV=/dev/xxxx # Only needed if running by hand
#SCRATCH_MNT=/scratch # Only needed if running by hand
here=`pwd`
tmp=/tmp/$$
status=1 # failure is the default!
_cleanup()
{
echo Cleanup beginning
rm $tmp.running
xfs_freeze -u $SCRATCH_MNT
sleep 10 # Give the dd loop time to finish
#Kill off any other possible stray stragglers that may be out there.
fuser -k -m /scratch_snap/dummy >/dev/null 2>&1
fuser -k -m /$SCRATCH_MNT/dummy >/dev/null 2>&1
wait
umount /scratch_snap
rmdir /scratch_snap
umount $SCRATCH_MNT
lvremove -f $VG/scratch_snap
lvremove -f $VG/scratch
#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 ]; then _notrun "This test requires that /scratch_snap not \
exist."; fi
mkdir /scratch_snap
#Verify we have the lvm user tools
LVM=`rpm -qa | grep 'lvm-'`
if [ x$LVM = "x" ]; then _notrun "LVM Must be installed to run this test"; fi
#Verify we have the a lvm enabled kernel
# TODO (This assumes lvm is a module. What if it is linked? I don't know how to \
check that. lsmod | grep lvm-mod;
if [ $? != 0 ]; then _notrun "This test requires the LVM kernel module be present"; \
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
# Create a large 64 Meg zero filled file on the LV
dd if=/dev/zero of=$SCRATCH_MNT/dummy bs=64k count=1000
#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 &
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
if [ $? != 0 ] ; then
echo mount for /scratch_snap failed
fi
umount /scratch_snap
if [ $? != 0 ] ; then
echo unmount for /scratch_snap failed
fi
lvremove -f $VG/scratch_snap > /dev/null 2>&1
if [ $? != 0 ] ; then
echo lvremove for /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.
_______________________________________________
linux-lvm mailing list
linux-lvm@sistina.com
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic