shaba/lvm2
1
0
Fork 0
mirror of git://sourceware.org/git/lvm2.git synced 2025-03-11 20:58:50 +03:00
Code Releases Activity
lvm2/test/shell/pvscan-autoactivate.sh

191 lines
4.5 KiB
Bash
Raw Normal View History

tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
#!/usr/bin/env bash
# Copyright (C) 2012 Red Hat, Inc. All rights reserved.
#
# This copyrighted material is made available to anyone wishing to use,
# modify, copy, or redistribute it subject to the terms and conditions
# of the GNU General Public License v.2.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software Foundation,
# Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
SKIP_WITH_LVMPOLLD=1
tests: add wait for udev Since the test is currently directly working with live directory, which can be getting updates from system's udev - add wait for settling so removal of all known PVs happens after that. But still this has major influce on behavior of running system, so the test should never be executed on a user used box.
2018-11-06 15:01:52 +01:00
RUNDIR="/run"
test -d "$RUNDIR" || RUNDIR="/var/run"
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
PVS_ONLINE_DIR="$RUNDIR/lvm/pvs_online"
VGS_ONLINE_DIR="$RUNDIR/lvm/vgs_online"
tests: add wait for udev Since the test is currently directly working with live directory, which can be getting updates from system's udev - add wait for settling so removal of all known PVs happens after that. But still this has major influce on behavior of running system, so the test should never be executed on a user used box.
2018-11-06 15:01:52 +01:00
# FIXME: kills logic for running system
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files() {
tests: add wait for udev Since the test is currently directly working with live directory, which can be getting updates from system's udev - add wait for settling so removal of all known PVs happens after that. But still this has major influce on behavior of running system, so the test should never be executed on a user used box.
2018-11-06 15:01:52 +01:00
# wait till udev is finished
aux udev_wait
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
rm -f "$PVS_ONLINE_DIR"/*
rm -f "$VGS_ONLINE_DIR"/*
tests: add wait for udev Since the test is currently directly working with live directory, which can be getting updates from system's udev - add wait for settling so removal of all known PVs happens after that. But still this has major influce on behavior of running system, so the test should never be executed on a user used box.
2018-11-06 15:01:52 +01:00
}
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
. lib/inittest
tests: check that pvscan --cache ignores certain PVs
2019-03-06 12:17:47 -06:00
aux prepare_pvs 3
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
vgcreate $vg1 "$dev1" "$dev2"
lvcreate -n $lv1 -l 4 -a n $vg1
tests: better support for /var/run
2018-10-16 23:46:50 +02:00
# the first pvscan scans all devs
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
test -d "$PVS_ONLINE_DIR" || mkdir -p "$PVS_ONLINE_DIR"
test -d "$VGS_ONLINE_DIR" || mkdir -p "$VGS_ONLINE_DIR"
_clear_online_files
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache -aay
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# the first pvscan scans all devs even when
# only one device is specified
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache -aay "$dev1"
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# touch foo to disable first-pvscan case,
# then check pvscan with no args scans all
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
touch "$RUNDIR/lvm/pvs_online/foo"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache -aay
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# touch foo to disable first-pvscan case,
# then check that vg is activated only after
# both devs appear separately
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
touch "$RUNDIR/lvm/pvs_online/foo"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
tests: fix shell quoting
2018-11-06 17:26:15 +01:00
pvscan --cache -aay "$dev1"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
check lv_field $vg1/$lv1 lv_active ""
tests: fix shell quoting
2018-11-06 17:26:15 +01:00
pvscan --cache -aay "$dev2"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# touch foo to disable first-pvscan case,
# then check that vg is activated when both
# devs appear together
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
touch "$RUNDIR/lvm/pvs_online/foo"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache -aay "$dev1" "$dev2"
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# Set up tests where one dev has no metadata
vgchange -an $vg1
vgremove -ff $vg1
pvremove "$dev1"
pvremove "$dev2"
pvcreate --metadatacopies 0 "$dev1"
pvcreate --metadatacopies 1 "$dev2"
vgcreate $vg1 "$dev1" "$dev2"
lvcreate -n $lv1 -l 4 -a n $vg1
# touch foo to disable first-pvscan case,
# test case where dev with metadata appears first
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
touch "$RUNDIR/lvm/pvs_online/foo"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache -aay "$dev2"
check lv_field $vg1/$lv1 lv_active ""
pvscan --cache -aay "$dev1"
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# touch foo to disable first-pvscan case,
# test case where dev without metadata
# appears first which triggers scanning all
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
touch "$RUNDIR/lvm/pvs_online/foo"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache -aay "$dev1"
check lv_field $vg1/$lv1 lv_active "active"
pvscan --cache -aay "$dev2"
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# dev without metadata is scanned, but
# first-pvscan case scans all devs
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache -aay "$dev1"
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# use the --cache option to record a dev
# is online without the -aay option to
# activate until after they are online
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
touch "$RUNDIR/lvm/pvs_online/foo"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache "$dev1"
check lv_field $vg1/$lv1 lv_active ""
pvscan --cache "$dev2"
check lv_field $vg1/$lv1 lv_active ""
pvscan --cache -aay
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
# like previous
pvscan: autoactivate a VG once When a VG has multiple PVs, and all those PVs come online at the same time, concurrent pvscans for each PV will all create the individual pvid files, and all will often see the VG is now complete. This causes each of the pvscan commands to think it should activate the VG, so there are multiple activations of the same VG. The vg lock serializes them, and only the first pvscan actually does the activation, but there is still a lot of extra overhead and time used by the other pvscans that attempt to activate the already active VG. This can lead to a backlog of pvscans and timeouts. To fix this, this adds a new /run/lvm/vgs_online/ dir that works like the existing /run/lvm/pvs_online/ dir. Each pvscan that wants to activate a VG will first try to exlusively create the file vgs_online/<vgname>. Only the first pvscan will succeed, and that one will do the VG activation. The other pvscans will find the vgname file exists and will not do the activation step. When a PV goes offline, the vgs_online file for the corresponding VG is removed. This allows the VG to be autoactivated again when the PV comes online again. This requires that the vgname be stored in the pvid files.
2019-02-13 14:21:56 -06:00
_clear_online_files
touch "$RUNDIR/lvm/pvs_online/foo"
tests: autoactivation tests for use without lvmetad Adjust a few lvmetad pvscan/autoactivation tests to be used without lvmetad, and add a test to cover some cases that have not been tested before.
2018-07-10 09:58:22 -05:00
pvscan --cache "$dev1"
check lv_field $vg1/$lv1 lv_active ""
pvscan --cache -aay "$dev2"
check lv_field $vg1/$lv1 lv_active "active"
lvchange -an $vg1
tests: check that pvscan --cache ignores certain PVs
2019-03-06 12:17:47 -06:00
# pvscan cache ignores pv that's not used
pvcreate "$dev3"
PVID3=`pvs $dev3 --noheading -o uuid | tr -d - | awk '{print $1}'`
echo $PVID3
not ls "$RUNDIR/lvm/pvs_online/$PVID3"
pvscan --cache -aay "$dev3"
ls "$RUNDIR/lvm/pvs_online"
not ls "$RUNDIR/lvm/pvs_online/$PVID3"
# pvscan cache ignores pv in a foreign vg
aux lvmconf "global/system_id_source = uname"
vgcreate $vg2 "$dev3"
lvcreate -an -n $lv1 -l1 $vg2
pvscan --cache -aay "$dev3"
ls "$RUNDIR/lvm/pvs_online/$PVID3"
check lv_field $vg2/$lv1 lv_active "active"
lvchange -an $vg2
rm "$RUNDIR/lvm/pvs_online/$PVID3"
vgchange -y --systemid "asdf" "$vg2"
pvscan --cache -aay "$dev3"
not ls "$RUNDIR/lvm/pvs_online/$PVID3"
check lv_field $vg2/$lv1 lv_active "" --foreign
Copy Permalink