Several SUSE bugfixes

Several SUSE bugfixes

[Thomas Renninger]

Based on version 043.
Would be great to see these mainline.

Harald, you removed udev rules for dmraid.
I expect adding all the kompat/partx udev rules may be a better approach.
But you better discuss this with Hannes and/or Neil.

Thanks a lot,

         Thomas

[PATCH 1/8] /etc/multipath.conf is not mandatory

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

If rootfs is on multipath, but platform does not have an /etc/multipath.conf
file which is not urgently needed, they system will not boot, due to:

multipathd is not started and rootfs and swap are not found:
systemctl status multipathd.service
* multipathd.service - Device-Mapper Multipath Device Controller
   Loaded: loaded (/usr/lib/systemd/system/multipathd.service; disabled; vendor preset: enabled)
   Active: inactive (dead)
Condition: start condition failed at Thu 2015-05-07 11:49:11 CEST; 7min ago
           ConditionPathExists=/etc/multipath.conf was not met
and exit to dracut shell.

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90multipath/multipathd.service | 1 -
 1 file changed, 1 deletion(-)

diff --git a/modules.d/90multipath/multipathd.service b/modules.d/90multipath/multipathd.service
index b64b02c..f7bc65f 100644
--- a/modules.d/90multipath/multipathd.service
+++ b/modules.d/90multipath/multipathd.service
@@ -6,7 +6,6 @@ Conflicts=shutdown.target
 ConditionKernelCommandLine=!nompath
 ConditionKernelCommandLine=!rd.multipath=0
 ConditionKernelCommandLine=!rd_NO_MULTIPATH
-ConditionPathExists=/etc/multipath.conf
 
 [Service]
 Type=simple
-- 
2.1.4

[PATCH 2/8] Fix nfs parsing when IPv4 address is used in /etc/fstab

[Thomas Renninger]

and remove an unneeded subprocess.

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95nfs/module-setup.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/modules.d/95nfs/module-setup.sh b/modules.d/95nfs/module-setup.sh
index 8eb542c..3f30285 100755
--- a/modules.d/95nfs/module-setup.sh
+++ b/modules.d/95nfs/module-setup.sh
@@ -51,9 +51,9 @@ cmdline() {
 
     ### ip= ###
     if [[ $nfs_device = [0-9]*\.[0-9]*\.[0-9]*.[0-9]* ]] || [[ $nfs_device = \[.*\] ]]; then
-        nfs_address="$nfs_device"
+        nfs_address="${nfs_device%%:*}"
     else
-        lookup=$(host $(echo ${nfs_device%%:*})| head -n1)
+        lookup=$(host "${nfs_device%%:*}"| head -n1)
         nfs_address=${lookup##* }
     fi
     ifname=$(ip -o route get to $nfs_address | sed -n 's/.*dev \([^ ]*\).*/\1/p')
-- 
2.1.4

[PATCH 3/8] Reset IFS variable

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

Setting and unsetting the IFS variable is tricky. To be on the
safe side we should always reset the IFS variable to its original
value after parsing.

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95fcoe/parse-fcoe.sh | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/modules.d/95fcoe/parse-fcoe.sh b/modules.d/95fcoe/parse-fcoe.sh
index b66fa3e..a214007 100755
--- a/modules.d/95fcoe/parse-fcoe.sh
+++ b/modules.d/95fcoe/parse-fcoe.sh
@@ -26,8 +26,10 @@ initqueue --onetime modprobe -b -q bnx2fc
 udevadm settle --timeout=30
 
 parse_fcoe_opts() {
+    local OLDIFS="$IFS"
     local IFS=:
     set $fcoe
+    IFS="$OLDIFS"
 
     case $# in
         2)
-- 
2.1.4

[PATCH 4/8] mkinitrd-suse.sh: Bail out with exit 1 if initrd cannot be generated

[Thomas Renninger]

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 mkinitrd-suse.sh | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/mkinitrd-suse.sh b/mkinitrd-suse.sh
index af42cbe..c2b314f 100755
--- a/mkinitrd-suse.sh
+++ b/mkinitrd-suse.sh
@@ -280,7 +280,8 @@ done
 
 [[ $targets && $kernels ]] || default_kernel_images
 if [[ ! $targets || ! $kernels ]];then
-    error "No kernel found in $boot_dir"
+    error "No kernel found in $boot_dir or bad modules dir in /lib/modules"
+    exit 1
 fi
 
 # We can have several targets/kernels, transform the list to an array
@@ -301,6 +302,8 @@ fi
 [[ $domu_module_list ]] || domu_module_list="${DOMU_INITRD_MODULES}"
 shopt -s extglob
 
+failed=""
+
 for ((i=0 ; $i<${#targets[@]} ; i++)); do
 
     if [[ $img_vers ]];then
@@ -329,14 +332,18 @@ for ((i=0 ; $i<${#targets[@]} ; i++)); do
         # expansion magics
         if [ -n "${modules_all}" ];then
             $dracut_cmd $dracut_args --force-drivers "${modules_all}" "$target" "$kernel" &>/dev/null
+            [ $? -ne 0 ] && failed="$failed $target"
         else
             $dracut_cmd $dracut_args "$target" "$kernel" &>/dev/null
+            [ $? -ne 0 ] && failed="$failed $target"
         fi
     else
         if [ -n "${modules_all}" ];then
             $dracut_cmd $dracut_args --force-drivers "${modules_all}" "$target" "$kernel"
+            [ $? -ne 0 ] && failed="$failed $target"
         else
             $dracut_cmd $dracut_args "$target" "$kernel"
+            [ $? -ne 0 ] && failed="$failed $target"
         fi
     fi
 done
@@ -345,4 +352,12 @@ if [ "$skip_update_bootloader" ] ; then
     echo 2>&1 "Did not refresh the bootloader. You might need to refresh it manually."
 else
     update-bootloader --refresh
+    [ $? -ne 0 ] && echo "Updating bootloader failed" && exit 1
+fi
+
+if [ "$failed" != "" ]; then
+    echo "Generating $failed targets failed"
+    exit 1
 fi
+
+exit 0
-- 
2.1.4

[PATCH 5/8] Use the uncompressed kernel file on arm and aarch to detect kernel version

[Thomas Renninger]

This is due to file cannot read out the kernel version on these.

Many thanks to Alexander Graf finding this in old SUSE mkinitrd code.
Converted by Thomas Renninger.

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>

https://bugzilla.suse.com/show_bug.cgi?id=908454
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 mkinitrd-suse.sh | 30 +++++++++++++++++++++++++++++-
 1 file changed, 29 insertions(+), 1 deletion(-)

diff --git a/mkinitrd-suse.sh b/mkinitrd-suse.sh
index c2b314f..0f2b3c4 100755
--- a/mkinitrd-suse.sh
+++ b/mkinitrd-suse.sh
@@ -145,6 +145,33 @@ is_xen_kernel() {
     return
 }
 
+# kernel_image_gz_from_image() and kernel_version_from_image() are helpers
+# for arm* kernels which produce zImage files which cannot be read from
+# get_kernel_version -> get rid of this workaround if possible
+kernel_image_gz_from_image() {
+    local arch=$(uname -i)
+    local r=${1}.gz
+
+    # uImage kernels can't be extracted directly. Use the vmlinux.gz instead
+    r=${r//uImage/vmlinux}
+
+    # on ARM a zImage can't be extracted directly. Other platforms define it
+    # as a gzipped vmlinux file, but not ARM. So only on ARM, use vmlinux.gz.
+    if [[ $arch =~ arm ]] || [[ $arch =~ aarch ]]; then
+        r=${r//zImage/vmlinux}
+    fi
+
+    echo $r
+}
+
+kernel_version_from_image() {
+    local kernel_image="$1" kernel_image_gz=$(kernel_image_gz_from_image "$1")
+
+    if get_kernel_version "$kernel_image" 2>/dev/null; then
+        return
+    fi
+    get_kernel_version "$kernel_image_gz" 2>/dev/null
+}
 
 # Taken over from SUSE mkinitrd
 default_kernel_images() {
@@ -183,7 +210,8 @@ default_kernel_images() {
 
         [ -L "$boot_dir/$kernel_image" ] && continue
         [ "${kernel_image%%.gz}" != "$kernel_image" ] && continue
-        kernel_version=$(/usr/bin/get_kernel_version \
+
+        kernel_version=$(kernel_version_from_image \
                          $boot_dir/$kernel_image 2> /dev/null)
         initrd_image=$(echo $kernel_image | sed -e "s|${regex}|initrd|")
         if [ "$kernel_image" != "$initrd_image" -a \
-- 
2.1.4

[PATCH 6/8] Install /etc/sysconfig/console to see specific fonts

[Thomas Renninger]

From: Werner Fink <werner-l3A5Bk7waGM@public.gmane.org>

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/10i18n/module-setup.sh | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/modules.d/10i18n/module-setup.sh b/modules.d/10i18n/module-setup.sh
index 9439875..398d02b 100755
--- a/modules.d/10i18n/module-setup.sh
+++ b/modules.d/10i18n/module-setup.sh
@@ -219,6 +219,10 @@ install() {
             print_vars KEYMAP EXT_KEYMAPS UNICODE FONT FONT_MAP FONT_UNIMAP >> ${initdir}${VCONFIG_CONF}
         fi
 
+        if dracut_module_included "systemd" && [[ -f /etc/sysconfig/console ]]; then
+            inst_simple /etc/sysconfig/console
+        fi
+
         return 0
     }
 
-- 
2.1.4

[PATCH 7/8] ensure pre-mount (and resume) run before root fsck

[Thomas Renninger]

From: Andrei Borzenkov <arvidjaar-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

References:
https://bugzilla.suse.com/show_bug.cgi?id=906592
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/98dracut-systemd/rootfs-generator.sh | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/modules.d/98dracut-systemd/rootfs-generator.sh b/modules.d/98dracut-systemd/rootfs-generator.sh
index 5910ce7..7eb0b1a 100755
--- a/modules.d/98dracut-systemd/rootfs-generator.sh
+++ b/modules.d/98dracut-systemd/rootfs-generator.sh
@@ -65,6 +65,23 @@ generator_mount_rootfs()
     fi
 }
 
+generator_fsck_after_pre_mount()
+{
+    local _name
+
+    [ -z "$1" ] && return 0
+
+    _name=$(dev_unit_name "$1")
+    [ -d /run/systemd/generator/systemd-fsck@${_name}.service.d ] || mkdir -p /run/systemd/generator/systemd-fsck@${_name}.service.d
+    if ! [ -f /run/systemd/generator/systemd-fsck@${_name}.service.d/after-pre-mount.conf ]; then
+        {
+            echo "[Unit]"
+            echo "After=dracut-pre-mount.service"
+        } > /run/systemd/generator/systemd-fsck@${_name}.service.d/after-pre-mount.conf
+    fi
+
+}
+
 root=$(getarg root=)
 case "$root" in
     block:LABEL=*|LABEL=*)
@@ -95,6 +112,7 @@ GENERATOR_DIR="$1"
 
 if [ "${root%%:*}" = "block" ]; then
    generator_wait_for_dev "${root#block:}" "$RDRETRY"
+   generator_fsck_after_pre_mount "${root#block:}"
    strstr "$(cat /proc/cmdline)" 'root=' || generator_mount_rootfs "${root#block:}" "$(getarg rootfstype=)" "$(getarg rootflags=)"
 fi
 
-- 
2.1.4

[PATCH 8/8] Use udev rules to create dmraid /dev/mapper/ devices

[Thomas Renninger]

From: Hannes Reinecke <hare-IBi9RG/b67k@public.gmane.org>

https://bugzilla.opensuse.org/show_bug.cgi?id=905746
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90dmraid/dmraid.sh       | 2 --
 modules.d/90dmraid/module-setup.sh | 2 ++
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/modules.d/90dmraid/dmraid.sh b/modules.d/90dmraid/dmraid.sh
index 3dcff38..cc4390f 100755
--- a/modules.d/90dmraid/dmraid.sh
+++ b/modules.d/90dmraid/dmraid.sh
@@ -26,8 +26,6 @@ if [ -n "$DM_RAIDS" ] || getargbool 0 rd.auto; then
                 if [ "${s##$r}" != "$s" ]; then
                     info "Activating $s"
                     dmraid -ay -i -p --rm_partitions "$s" 2>&1 | vinfo
-                    [ -e "/dev/mapper/$s" ] && kpartx -a "/dev/mapper/$s" 2>&1 | vinfo
-                    udevsettle
                 fi
             done
         done
diff --git a/modules.d/90dmraid/module-setup.sh b/modules.d/90dmraid/module-setup.sh
index e8de5f5..b77bbcf 100755
--- a/modules.d/90dmraid/module-setup.sh
+++ b/modules.d/90dmraid/module-setup.sh
@@ -74,6 +74,8 @@ install() {
 
     inst "$moddir/dmraid.sh" /sbin/dmraid_scan
 
+    inst_rules 64-md-raid.rules 66-kpartx.rules 67-kpartx-compat.rules
+
     inst_libdir_file "libdmraid-events*.so*"
 
     inst_rules "$moddir/61-dmraid-imsm.rules"
-- 
2.1.4

Re: [PATCH 6/8] Install /etc/sysconfig/console to see specific fonts

[Andrei Borzenkov]

В Fri, 26 Jun 2015 16:55:55 +0200
Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org> пишет:

> From: Werner Fink <werner-l3A5Bk7waGM@public.gmane.org>
> 
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/10i18n/module-setup.sh | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/modules.d/10i18n/module-setup.sh b/modules.d/10i18n/module-setup.sh
> index 9439875..398d02b 100755
> --- a/modules.d/10i18n/module-setup.sh
> +++ b/modules.d/10i18n/module-setup.sh
> @@ -219,6 +219,10 @@ install() {
>              print_vars KEYMAP EXT_KEYMAPS UNICODE FONT FONT_MAP FONT_UNIMAP >> ${initdir}${VCONFIG_CONF}
>          fi
>  
> +        if dracut_module_included "systemd" && [[ -f /etc/sysconfig/console ]]; then
> +            inst_simple /etc/sysconfig/console
> +        fi
> +
>          return 0
>      }
>  

Upstream systemd does not even look at this file. Could you explain
what problem is fixed here and why presence of this file fixes it?

Re: [PATCH 8/8] Use udev rules to create dmraid /dev/mapper/ devices

[Andrei Borzenkov]

В Fri, 26 Jun 2015 16:55:57 +0200
Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org> пишет:

> From: Hannes Reinecke <hare-IBi9RG/b67k@public.gmane.org>
> 
> https://bugzilla.opensuse.org/show_bug.cgi?id=905746
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/90dmraid/dmraid.sh       | 2 --
>  modules.d/90dmraid/module-setup.sh | 2 ++
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/modules.d/90dmraid/dmraid.sh b/modules.d/90dmraid/dmraid.sh
> index 3dcff38..cc4390f 100755
> --- a/modules.d/90dmraid/dmraid.sh
> +++ b/modules.d/90dmraid/dmraid.sh
> @@ -26,8 +26,6 @@ if [ -n "$DM_RAIDS" ] || getargbool 0 rd.auto; then
>                  if [ "${s##$r}" != "$s" ]; then
>                      info "Activating $s"
>                      dmraid -ay -i -p --rm_partitions "$s" 2>&1 | vinfo
> -                    [ -e "/dev/mapper/$s" ] && kpartx -a "/dev/mapper/$s" 2>&1 | vinfo
> -                    udevsettle
>                  fi
>              done
>          done
> diff --git a/modules.d/90dmraid/module-setup.sh b/modules.d/90dmraid/module-setup.sh
> index e8de5f5..b77bbcf 100755
> --- a/modules.d/90dmraid/module-setup.sh
> +++ b/modules.d/90dmraid/module-setup.sh
> @@ -74,6 +74,8 @@ install() {
>  
>      inst "$moddir/dmraid.sh" /sbin/dmraid_scan
>  
> +    inst_rules 64-md-raid.rules 66-kpartx.rules 67-kpartx-compat.rules
> +

Why 64-md-raid.rules is here? 

>      inst_libdir_file "libdmraid-events*.so*"
>  
>      inst_rules "$moddir/61-dmraid-imsm.rules"

Re: [PATCH 8/8] Use udev rules to create dmraid /dev/mapper/ devices

[Thomas Renninger]

On Saturday, June 27, 2015 08:41:53 AM Andrei Borzenkov wrote:
> Ð’ Fri, 26 Jun 2015 16:55:57 +0200
> 
> > diff --git a/modules.d/90dmraid/module-setup.sh
> > b/modules.d/90dmraid/module-setup.sh index e8de5f5..b77bbcf 100755
> > --- a/modules.d/90dmraid/module-setup.sh
> > +++ b/modules.d/90dmraid/module-setup.sh
> > @@ -74,6 +74,8 @@ install() {
> > 
> >      inst "$moddir/dmraid.sh" /sbin/dmraid_scan
> > 
> > +    inst_rules 64-md-raid.rules 66-kpartx.rules 67-kpartx-compat.rules
> > +
> 
> Why 64-md-raid.rules is here?

Probably the same reason why it slipped in once:
I mixed up md and dm and thought it is a dmraid rule which should
get in again if kpartx rules are used.

Thanks for finding this, the rule is already added at the right place:
modules.d/90mdraid/module-setup.sh
and should not show up here again.

I repost this one.

     Thomas

[PATCH] Use udev rules to create dmraid /dev/mapper/ devices

[Thomas Renninger]

From: Hannes Reinecke <hare-IBi9RG/b67k@public.gmane.org>

https://bugzilla.opensuse.org/show_bug.cgi?id=905746

Version 2: Remove 64-md-raid.rules

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90dmraid/dmraid.sh       | 2 --
 modules.d/90dmraid/module-setup.sh | 2 ++
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/modules.d/90dmraid/dmraid.sh b/modules.d/90dmraid/dmraid.sh
index 3dcff38..cc4390f 100755
--- a/modules.d/90dmraid/dmraid.sh
+++ b/modules.d/90dmraid/dmraid.sh
@@ -26,8 +26,6 @@ if [ -n "$DM_RAIDS" ] || getargbool 0 rd.auto; then
                 if [ "${s##$r}" != "$s" ]; then
                     info "Activating $s"
                     dmraid -ay -i -p --rm_partitions "$s" 2>&1 | vinfo
-                    [ -e "/dev/mapper/$s" ] && kpartx -a "/dev/mapper/$s" 2>&1 | vinfo
-                    udevsettle
                 fi
             done
         done
diff --git a/modules.d/90dmraid/module-setup.sh b/modules.d/90dmraid/module-setup.sh
index e8de5f5..797a58e 100755
--- a/modules.d/90dmraid/module-setup.sh
+++ b/modules.d/90dmraid/module-setup.sh
@@ -74,6 +74,8 @@ install() {
 
     inst "$moddir/dmraid.sh" /sbin/dmraid_scan
 
+    inst_rules 66-kpartx.rules 67-kpartx-compat.rules
+
     inst_libdir_file "libdmraid-events*.so*"
 
     inst_rules "$moddir/61-dmraid-imsm.rules"
-- 
2.1.4

Re: [PATCH] Use udev rules to create dmraid /dev/mapper/ devices

[Dracut GitHub Import Bot]

Patchset imported to github.
Pull request:
<https://github.com/haraldh/dracut/compare/master...dracut-mailing-devs:1435599415-10187-1-git-send-email-trenn%40suse.de>

Re: [PATCH 7/8] ensure pre-mount (and resume) run before root fsck

[Alexander Tsoy]

В Fri, 26 Jun 2015 16:55:56 +0200
Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org> пишет:

> From: Andrei Borzenkov <arvidjaar-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> 
> References:
> https://bugzilla.suse.com/show_bug.cgi?id=906592
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/98dracut-systemd/rootfs-generator.sh | 18
> ++++++++++++++++++ 1 file changed, 18 insertions(+)
> 
> diff --git a/modules.d/98dracut-systemd/rootfs-generator.sh
> b/modules.d/98dracut-systemd/rootfs-generator.sh index
> 5910ce7..7eb0b1a 100755 ---
> a/modules.d/98dracut-systemd/rootfs-generator.sh +++
> b/modules.d/98dracut-systemd/rootfs-generator.sh @@ -65,6 +65,23 @@
> generator_mount_rootfs() fi
>  }
>  
> +generator_fsck_after_pre_mount()
> +{
> +    local _name
> +
> +    [ -z "$1" ] && return 0
> +
> +    _name=$(dev_unit_name "$1")
> +    [ -d /run/systemd/generator/systemd-fsck@${_name}.service.d ] ||
> mkdir -p /run/systemd/generator/systemd-fsck@${_name}.service.d
> +    if !
> [ -f /run/systemd/generator/systemd-fsck@${_name}.service.d/after-pre-mount.conf
> ]; then
> +        {
> +            echo "[Unit]"
> +            echo "After=dracut-pre-mount.service"
> +        }
> > /run/systemd/generator/systemd-fsck@${_name}.service.d/after-pre-mount.conf
> +    fi
> +
> +}
> +
>  root=$(getarg root=)
>  case "$root" in
>      block:LABEL=*|LABEL=*)
> @@ -95,6 +112,7 @@ GENERATOR_DIR="$1"
>  
>  if [ "${root%%:*}" = "block" ]; then
>     generator_wait_for_dev "${root#block:}" "$RDRETRY"
> +   generator_fsck_after_pre_mount "${root#block:}"
>     strstr "$(cat /proc/cmdline)" 'root=' || generator_mount_rootfs
> "${root#block:}" "$(getarg rootfstype=)" "$(getarg rootflags=)" fi
>  

Hmmm.. How about systemd >=220? See the following change:
http://cgit.freedesktop.org/systemd/systemd/commit/?id=4dda4e637e4c17a14db6cd265f36f5e8a5050367

Shouldn't we also add "Before=systemd-fsck-root.service" to
dracut-pre-mount.service?

-- 
Alexander Tsoy

Re: [PATCH 1/8] /etc/multipath.conf is not mandatory

[Harald Hoyer]

On 26.06.2015 16:55, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> 
> If rootfs is on multipath, but platform does not have an /etc/multipath.conf
> file which is not urgently needed, they system will not boot, due to:
> 
> multipathd is not started and rootfs and swap are not found:
> systemctl status multipathd.service
> * multipathd.service - Device-Mapper Multipath Device Controller
>    Loaded: loaded (/usr/lib/systemd/system/multipathd.service; disabled; vendor preset: enabled)
>    Active: inactive (dead)
> Condition: start condition failed at Thu 2015-05-07 11:49:11 CEST; 7min ago
>            ConditionPathExists=/etc/multipath.conf was not met
> and exit to dracut shell.
> 
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/90multipath/multipathd.service | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/modules.d/90multipath/multipathd.service b/modules.d/90multipath/multipathd.service
> index b64b02c..f7bc65f 100644
> --- a/modules.d/90multipath/multipathd.service
> +++ b/modules.d/90multipath/multipathd.service
> @@ -6,7 +6,6 @@ Conflicts=shutdown.target
>  ConditionKernelCommandLine=!nompath
>  ConditionKernelCommandLine=!rd.multipath=0
>  ConditionKernelCommandLine=!rd_NO_MULTIPATH
> -ConditionPathExists=/etc/multipath.conf
>  
>  [Service]
>  Type=simple
> 

CC'ing Benjmain Marzinski as the Fedora maintainer of device-mapper-multipath.

I don't understand how multipath is used without a configuration?

Anyway, I see the service file in Fedora is quiet different from yours and
yours is also different from upstream.

So we have:

*Upstream* -
<http://git.opensvc.com/gitweb.cgi?p=multipath-tools/.git;a=blob_plain;f=multipathd/multipathd.service;h=b5b755bbacd08cd9f612206ce73dce965dbc8b2a;hb=HEAD>

[Unit]
Description=Device-Mapper Multipath Device Controller
Before=iscsi.service iscsid.service lvm2-activation-early.service
Before=local-fs-pre.target
After=multipathd.socket
DefaultDependencies=no
Wants=local-fs-pre.target multipathd.socket
Conflicts=shutdown.target

[Service]
Type=notify
NotifyAccess=main
LimitCORE=infinity
ExecStartPre=/sbin/modprobe dm-multipath
ExecStart=/sbin/multipathd -d -s
ExecReload=/sbin/multipathd reconfigure

[…]

*Fedora device-mapper-multipath-0.4.9-76*

 [Unit]
 Description=Device-Mapper Multipath Device Controller
 Before=iscsi.service iscsid.service lvm2-activation-early.service
-Before=local-fs-pre.target
-After=multipathd.socket
+After=syslog.target
+ConditionPathExists=/etc/multipath.conf
+ConditionKernelCommandLine=!nompath
 DefaultDependencies=no
-Wants=local-fs-pre.target multipathd.socket
 Conflicts=shutdown.target

 [Service]
-Type=notify
-NotifyAccess=main
-LimitCORE=infinity
+Type=forking
+PIDFile=/var/run/multipathd/multipathd.pid
 ExecStartPre=/sbin/modprobe dm-multipath
-ExecStart=/sbin/multipathd -d -s
+ExecStartPre=-/sbin/multipath -A
+ExecStart=/sbin/multipathd
 ExecReload=/sbin/multipathd reconfigure
+#ExecStop=/path/to/scrip delete-me if not necessary


*openSUSE multipath-tools-0.5.0-20.1*

 [Unit]
 Description=Device-Mapper Multipath Device Controller
 Before=iscsi.service iscsid.service lvm2-activation-early.service
-Before=local-fs-pre.target
-After=multipathd.socket
+Before=local-fs-pre.target systemd-udev-trigger.service
+After=multipathd.socket systemd-udevd.service
 DefaultDependencies=no
 Wants=local-fs-pre.target multipathd.socket
 Conflicts=shutdown.target
@@ -11,6 +11,10 @@
 Type=notify
 NotifyAccess=main
 LimitCORE=infinity
-ExecStartPre=/sbin/modprobe dm-multipath
+ExecStartPre=/sbin/modprobe -a scsi_dh_alua scsi_dh_emc scsi_dh_rdac dm-multipath
 ExecStart=/sbin/multipathd -d -s
 ExecReload=/sbin/multipathd reconfigure



I think it's impossible to make this independent in dracut from the shipped
package. You guys definetly should either ship the initramfs version in the
package, so the dracut module can copy that, or better use the same unit file
across distributions.

Re: [PATCH 7/8] ensure pre-mount (and resume) run before root fsck

[Harald Hoyer]

On 30.06.2015 01:49, Alexander Tsoy wrote:
> В Fri, 26 Jun 2015 16:55:56 +0200
> Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org> пишет:
> 
>> From: Andrei Borzenkov <arvidjaar-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>>
>> References:
>> https://bugzilla.suse.com/show_bug.cgi?id=906592
>> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
>> ---
>>  modules.d/98dracut-systemd/rootfs-generator.sh | 18
>> ++++++++++++++++++ 1 file changed, 18 insertions(+)
>>
>> diff --git a/modules.d/98dracut-systemd/rootfs-generator.sh
>> b/modules.d/98dracut-systemd/rootfs-generator.sh index
>> 5910ce7..7eb0b1a 100755 ---
>> a/modules.d/98dracut-systemd/rootfs-generator.sh +++
>> b/modules.d/98dracut-systemd/rootfs-generator.sh @@ -65,6 +65,23 @@
>> generator_mount_rootfs() fi
>>  }
>>  
>> +generator_fsck_after_pre_mount()
>> +{
>> +    local _name
>> +
>> +    [ -z "$1" ] && return 0
>> +
>> +    _name=$(dev_unit_name "$1")
>> +    [ -d /run/systemd/generator/systemd-fsck@${_name}.service.d ] ||
>> mkdir -p /run/systemd/generator/systemd-fsck@${_name}.service.d
>> +    if !
>> [ -f /run/systemd/generator/systemd-fsck@${_name}.service.d/after-pre-mount.conf
>> ]; then
>> +        {
>> +            echo "[Unit]"
>> +            echo "After=dracut-pre-mount.service"
>> +        }
>>> /run/systemd/generator/systemd-fsck@${_name}.service.d/after-pre-mount.conf
>> +    fi
>> +
>> +}
>> +
>>  root=$(getarg root=)
>>  case "$root" in
>>      block:LABEL=*|LABEL=*)
>> @@ -95,6 +112,7 @@ GENERATOR_DIR="$1"
>>  
>>  if [ "${root%%:*}" = "block" ]; then
>>     generator_wait_for_dev "${root#block:}" "$RDRETRY"
>> +   generator_fsck_after_pre_mount "${root#block:}"
>>     strstr "$(cat /proc/cmdline)" 'root=' || generator_mount_rootfs
>> "${root#block:}" "$(getarg rootfstype=)" "$(getarg rootflags=)" fi
>>  
> 
> Hmmm.. How about systemd >=220? See the following change:
> http://cgit.freedesktop.org/systemd/systemd/commit/?id=4dda4e637e4c17a14db6cd265f36f5e8a5050367
> 
> Shouldn't we also add "Before=systemd-fsck-root.service" to
> dracut-pre-mount.service?
> 

Good point.. yes

Re: [PATCH] Use udev rules to create dmraid /dev/mapper/ devices

[Harald Hoyer]

On 29.06.2015 19:36, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-IBi9RG/b67k@public.gmane.org>
> 
> https://bugzilla.opensuse.org/show_bug.cgi?id=905746
> 
> Version 2: Remove 64-md-raid.rules
> 
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/90dmraid/dmraid.sh       | 2 --
>  modules.d/90dmraid/module-setup.sh | 2 ++
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/modules.d/90dmraid/dmraid.sh b/modules.d/90dmraid/dmraid.sh
> index 3dcff38..cc4390f 100755
> --- a/modules.d/90dmraid/dmraid.sh
> +++ b/modules.d/90dmraid/dmraid.sh
> @@ -26,8 +26,6 @@ if [ -n "$DM_RAIDS" ] || getargbool 0 rd.auto; then
>                  if [ "${s##$r}" != "$s" ]; then
>                      info "Activating $s"
>                      dmraid -ay -i -p --rm_partitions "$s" 2>&1 | vinfo
> -                    [ -e "/dev/mapper/$s" ] && kpartx -a "/dev/mapper/$s" 2>&1 | vinfo
> -                    udevsettle
>                  fi
>              done
>          done
> diff --git a/modules.d/90dmraid/module-setup.sh b/modules.d/90dmraid/module-setup.sh
> index e8de5f5..797a58e 100755
> --- a/modules.d/90dmraid/module-setup.sh
> +++ b/modules.d/90dmraid/module-setup.sh
> @@ -74,6 +74,8 @@ install() {
>  
>      inst "$moddir/dmraid.sh" /sbin/dmraid_scan
>  
> +    inst_rules 66-kpartx.rules 67-kpartx-compat.rules
> +
>      inst_libdir_file "libdmraid-events*.so*"
>  
>      inst_rules "$moddir/61-dmraid-imsm.rules"
> 

Fedora does not have 66-kpartx.rules nor 67-kpartx-compat.rules ...

Heinz, do we need the kpartx part still?

I reverted to kpartx, because "dmraid" adds a "p" as a seperator by default for
the partitions and this breaks existing installations.

Re: [PATCH] Use udev rules to create dmraid /dev/mapper/ devices

[Hannes Reinecke]

On 07/07/2015 01:41 PM, Harald Hoyer wrote:
> On 29.06.2015 19:36, Thomas Renninger wrote:
>> From: Hannes Reinecke <hare-IBi9RG/b67k@public.gmane.org>
>>
>> https://bugzilla.opensuse.org/show_bug.cgi?id=905746
>>
>> Version 2: Remove 64-md-raid.rules
>>
>> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
>> ---
>>  modules.d/90dmraid/dmraid.sh       | 2 --
>>  modules.d/90dmraid/module-setup.sh | 2 ++
>>  2 files changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/modules.d/90dmraid/dmraid.sh b/modules.d/90dmraid/dmraid.sh
>> index 3dcff38..cc4390f 100755
>> --- a/modules.d/90dmraid/dmraid.sh
>> +++ b/modules.d/90dmraid/dmraid.sh
>> @@ -26,8 +26,6 @@ if [ -n "$DM_RAIDS" ] || getargbool 0 rd.auto; then
>>                  if [ "${s##$r}" != "$s" ]; then
>>                      info "Activating $s"
>>                      dmraid -ay -i -p --rm_partitions "$s" 2>&1 | vinfo
>> -                    [ -e "/dev/mapper/$s" ] && kpartx -a "/dev/mapper/$s" 2>&1 | vinfo
>> -                    udevsettle
>>                  fi
>>              done
>>          done
>> diff --git a/modules.d/90dmraid/module-setup.sh b/modules.d/90dmraid/module-setup.sh
>> index e8de5f5..797a58e 100755
>> --- a/modules.d/90dmraid/module-setup.sh
>> +++ b/modules.d/90dmraid/module-setup.sh
>> @@ -74,6 +74,8 @@ install() {
>>  
>>      inst "$moddir/dmraid.sh" /sbin/dmraid_scan
>>  
>> +    inst_rules 66-kpartx.rules 67-kpartx-compat.rules
>> +
>>      inst_libdir_file "libdmraid-events*.so*"
>>  
>>      inst_rules "$moddir/61-dmraid-imsm.rules"
>>
> 
> Fedora does not have 66-kpartx.rules nor 67-kpartx-compat.rules ...
> 
> Heinz, do we need the kpartx part still?
> 
> I reverted to kpartx, because "dmraid" adds a "p" as a seperator by default for
> the partitions and this breaks existing installations.
> 

I would recommend splitting the kpartx call into a separate udev
rule; otherwise you'll run into timing issues with udev.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		               zSeries & Storage
hare-IBi9RG/b67k@public.gmane.org			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)

Re: [PATCH] Use udev rules to create dmraid /dev/mapper/ devices

[Harald Hoyer]

On 07.07.2015 13:54, Hannes Reinecke wrote:
> On 07/07/2015 01:41 PM, Harald Hoyer wrote:
>> On 29.06.2015 19:36, Thomas Renninger wrote:
>>> From: Hannes Reinecke <hare-IBi9RG/b67k@public.gmane.org>
>>>
>>> https://bugzilla.opensuse.org/show_bug.cgi?id=905746
>>>
>>> Version 2: Remove 64-md-raid.rules
>>>
>>> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
>>> ---
>>>  modules.d/90dmraid/dmraid.sh       | 2 --
>>>  modules.d/90dmraid/module-setup.sh | 2 ++
>>>  2 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/modules.d/90dmraid/dmraid.sh b/modules.d/90dmraid/dmraid.sh
>>> index 3dcff38..cc4390f 100755
>>> --- a/modules.d/90dmraid/dmraid.sh
>>> +++ b/modules.d/90dmraid/dmraid.sh
>>> @@ -26,8 +26,6 @@ if [ -n "$DM_RAIDS" ] || getargbool 0 rd.auto; then
>>>                  if [ "${s##$r}" != "$s" ]; then
>>>                      info "Activating $s"
>>>                      dmraid -ay -i -p --rm_partitions "$s" 2>&1 | vinfo
>>> -                    [ -e "/dev/mapper/$s" ] && kpartx -a "/dev/mapper/$s" 2>&1 | vinfo
>>> -                    udevsettle
>>>                  fi
>>>              done
>>>          done
>>> diff --git a/modules.d/90dmraid/module-setup.sh b/modules.d/90dmraid/module-setup.sh
>>> index e8de5f5..797a58e 100755
>>> --- a/modules.d/90dmraid/module-setup.sh
>>> +++ b/modules.d/90dmraid/module-setup.sh
>>> @@ -74,6 +74,8 @@ install() {
>>>  
>>>      inst "$moddir/dmraid.sh" /sbin/dmraid_scan
>>>  
>>> +    inst_rules 66-kpartx.rules 67-kpartx-compat.rules
>>> +
>>>      inst_libdir_file "libdmraid-events*.so*"
>>>  
>>>      inst_rules "$moddir/61-dmraid-imsm.rules"
>>>
>>
>> Fedora does not have 66-kpartx.rules nor 67-kpartx-compat.rules ...
>>
>> Heinz, do we need the kpartx part still?
>>
>> I reverted to kpartx, because "dmraid" adds a "p" as a seperator by default for
>> the partitions and this breaks existing installations.
>>
> 
> I would recommend splitting the kpartx call into a separate udev
> rule; otherwise you'll run into timing issues with udev.
> 
> Cheers,
> 
> Hannes
> 

Care to share your 66-kpartx.rules 67-kpartx-compat.rules or make them
upstream? Shouldn't we agree on one naming scheme?

Re: [PATCH] Use udev rules to create dmraid /dev/mapper/ devices

[Hannes Reinecke]

On 11/13/2015 12:00 PM, Harald Hoyer wrote:
> On 07.07.2015 13:54, Hannes Reinecke wrote:
>> On 07/07/2015 01:41 PM, Harald Hoyer wrote:
>>> On 29.06.2015 19:36, Thomas Renninger wrote:
>>>> From: Hannes Reinecke <hare-IBi9RG/b67k@public.gmane.org>
>>>>
>>>> https://bugzilla.opensuse.org/show_bug.cgi?id=905746
>>>>
>>>> Version 2: Remove 64-md-raid.rules
>>>>
>>>> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
>>>> ---
>>>>  modules.d/90dmraid/dmraid.sh       | 2 --
>>>>  modules.d/90dmraid/module-setup.sh | 2 ++
>>>>  2 files changed, 2 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/modules.d/90dmraid/dmraid.sh b/modules.d/90dmraid/dmraid.sh
>>>> index 3dcff38..cc4390f 100755
>>>> --- a/modules.d/90dmraid/dmraid.sh
>>>> +++ b/modules.d/90dmraid/dmraid.sh
>>>> @@ -26,8 +26,6 @@ if [ -n "$DM_RAIDS" ] || getargbool 0 rd.auto; then
>>>>                  if [ "${s##$r}" != "$s" ]; then
>>>>                      info "Activating $s"
>>>>                      dmraid -ay -i -p --rm_partitions "$s" 2>&1 | vinfo
>>>> -                    [ -e "/dev/mapper/$s" ] && kpartx -a "/dev/mapper/$s" 2>&1 | vinfo
>>>> -                    udevsettle
>>>>                  fi
>>>>              done
>>>>          done
>>>> diff --git a/modules.d/90dmraid/module-setup.sh b/modules.d/90dmraid/module-setup.sh
>>>> index e8de5f5..797a58e 100755
>>>> --- a/modules.d/90dmraid/module-setup.sh
>>>> +++ b/modules.d/90dmraid/module-setup.sh
>>>> @@ -74,6 +74,8 @@ install() {
>>>>  
>>>>      inst "$moddir/dmraid.sh" /sbin/dmraid_scan
>>>>  
>>>> +    inst_rules 66-kpartx.rules 67-kpartx-compat.rules
>>>> +
>>>>      inst_libdir_file "libdmraid-events*.so*"
>>>>  
>>>>      inst_rules "$moddir/61-dmraid-imsm.rules"
>>>>
>>>
>>> Fedora does not have 66-kpartx.rules nor 67-kpartx-compat.rules ...
>>>
>>> Heinz, do we need the kpartx part still?
>>>
>>> I reverted to kpartx, because "dmraid" adds a "p" as a seperator by default for
>>> the partitions and this breaks existing installations.
>>>
>>
>> I would recommend splitting the kpartx call into a separate udev
>> rule; otherwise you'll run into timing issues with udev.
>>
>> Cheers,
>>
>> Hannes
>>
> 
> Care to share your 66-kpartx.rules 67-kpartx-compat.rules or make them
> upstream? Shouldn't we agree on one naming scheme?
> 
Sure, I'd love to.

Although the 67-kpartx-compat.rules is pretty much SUSE specific, as
we've messed up the device-mapper device naming
(older SUSE versions would be creating device-mapper names with
underscores, but the links would be using normal dashes).
So it's not relevant to the general audience, but yeah I'd like
to have them both included.

Will be sending patches to multipath-tools covering that.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		               zSeries & Storage
hare-IBi9RG/b67k@public.gmane.org			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)