[PATCH] livecd-iso-to-disk --format fails with existing LV on USB

[David Huff]

Resolves rhbz#583658
---
 tools/livecd-iso-to-disk.sh |   16 ++++++++++++++++
 1 files changed, 16 insertions(+), 0 deletions(-)

diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh
index 3a4726d..0cfcfb8 100755
--- a/tools/livecd-iso-to-disk.sh
+++ b/tools/livecd-iso-to-disk.sh
@@ -138,6 +138,21 @@ checkPartActive() {
     fi
 }
 
+checkLVMActive() {
+    dev=$1
+
+	# check for an active lv which will cause format to fail
+	for vg in $(pvs -o vg_name --noheadings $dev $dev${dev_delim}[0-9]* 2>/dev/null|sort
-u); do
+		if [ $(lvs -o lv_attr --noheadings $vg |awk {'print $1;'} |cut -c 5) =
"a" ]; then
+		echo "Device, $dev, contains a volume group with an active volume and cannot be
formated!"
+		echo "You can remove the volume group with "
+		echo "    # /sbin/vgremove -f $vg"
+		exitclean
+		fi
+    done
+    return 0
+}
+
 createGPTLayout() {
     dev=$1
     getdisk $dev
@@ -407,6 +422,7 @@ fi
 # do some basic sanity checks.  
 checkMounted $USBDEV
 if [ -n "$format" ];then
+  checkLVMActive $USBDEV
   # checks for a valid filesystem
   if [ -n "$efi" ];then
     createGPTLayout $USBDEV
-- 
1.6.6.1

[Sebastian Dziallas]

I'm going to make some noise here as this is already the second patch
that has hit the list which needs a reaction - I don't have commit
access to that repo, but Frederick [1] [2] had also posted another one
earlier. I'm CC'ing a couple of folks that *do* have commit access
according to FAS. It'd be awesome if we could find a solution how to
proceed with the livecd-tools in general, as it seems pretty dead
right now.

--Sebastian

[1] http://lists.fedoraproject.org/pipermail/livecd/2010-April/005764.html
[2] http://lists.fedoraproject.org/pipermail/livecd/2010-March/005762.html

On Thu, Apr 22, 2010 at 7:03 PM, David Huff <dhuff(a)redhat.com&gt; wrote:
>  Resolves rhbz#583658
>  ---
>   tools/livecd-iso-to-disk.sh |   16 ++++++++++++++++
>   1 files changed, 16 insertions(+), 0 deletions(-)
>
>  diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh
>  index 3a4726d..0cfcfb8 100755
>  --- a/tools/livecd-iso-to-disk.sh
>  +++ b/tools/livecd-iso-to-disk.sh
>  @@ -138,6 +138,21 @@ checkPartActive() {
>      fi
>   }
>
>  +checkLVMActive() {
>  +    dev=$1
>  +
>  +   # check for an active lv which will cause format to fail
>  +   for vg in $(pvs -o vg_name --noheadings $dev $dev${dev_delim}[0-9]*
> 2>/dev/null|sort -u); do
>  +       if [ $(lvs -o lv_attr --noheadings $vg |awk {'print $1;'} |cut -c 5) =
> "a" ]; then
>  +       echo "Device, $dev, contains a volume group with an active volume and cannot
> be formated!"
>  +       echo "You can remove the volume group with "
>  +       echo "    # /sbin/vgremove -f $vg"
>  +       exitclean
>  +       fi
>  +    done
>  +    return 0
>  +}
>  +
>   createGPTLayout() {
>      dev=$1
>      getdisk $dev
>  @@ -407,6 +422,7 @@ fi
>   # do some basic sanity checks.
>   checkMounted $USBDEV
>   if [ -n "$format" ];then
>  +  checkLVMActive $USBDEV
>    # checks for a valid filesystem
>    if [ -n "$efi" ];then
>      createGPTLayout $USBDEV
>  --
>  1.6.6.1
>
>  --
>  livecd mailing list
>  livecd(a)lists.fedoraproject.org
>  https://admin.fedoraproject.org/mailman/listinfo/livecd

[James Heather]

I have also been wondering whether livecd-tools is being actively
maintained.

I reported this

https://bugzilla.redhat.com/show_bug.cgi?id=542167

back in November, and so far there's no comment from any of the
maintainers. It is an important bug, and pretty trivial to fix for
someone who knows Python (unlike me).

James

On Thu, 2010-04-22 at 22:00 +0100, Sebastian Dziallas wrote:

>  I'm going to make some noise here as this is already the second
> patch
>  that has hit the list which needs a reaction - I don't have commit
>  access to that repo, but Frederick [1] [2] had also posted another one
>  earlier. I'm CC'ing a couple of folks that *do* have commit access
>  according to FAS. It'd be awesome if we could find a solution how to
>  proceed with the livecd-tools in general, as it seems pretty dead
>  right now.
>  
>  --Sebastian
>  
>  [1] http://lists.fedoraproject.org/pipermail/livecd/2010-April/005764.html
>  [2] http://lists.fedoraproject.org/pipermail/livecd/2010-March/005762.html
>  
>  On Thu, Apr 22, 2010 at 7:03 PM, David Huff <dhuff(a)redhat.com&gt; wrote:
>  > Resolves rhbz#583658
>  > ---
>  >  tools/livecd-iso-to-disk.sh |   16 ++++++++++++++++
>  >  1 files changed, 16 insertions(+), 0 deletions(-)
>  >
>  > diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh
>  > index 3a4726d..0cfcfb8 100755
>  > --- a/tools/livecd-iso-to-disk.sh
>  > +++ b/tools/livecd-iso-to-disk.sh
>  > @@ -138,6 +138,21 @@ checkPartActive() {
>  >     fi
>  >  }
>  >
>  > +checkLVMActive() {
>  > +    dev=$1
>  > +
>  > +   # check for an active lv which will cause format to fail
>  > +   for vg in $(pvs -o vg_name --noheadings $dev $dev${dev_delim}[0-9]*
> 2>/dev/null|sort -u); do
>  > +       if [ $(lvs -o lv_attr --noheadings $vg |awk {'print $1;'} |cut -c 5)
> = "a" ]; then
>  > +       echo "Device, $dev, contains a volume group with an active volume and
> cannot be formated!"
>  > +       echo "You can remove the volume group with "
>  > +       echo "    # /sbin/vgremove -f $vg"
>  > +       exitclean
>  > +       fi
>  > +    done
>  > +    return 0
>  > +}
>  > +
>  >  createGPTLayout() {
>  >     dev=$1
>  >     getdisk $dev
>  > @@ -407,6 +422,7 @@ fi
>  >  # do some basic sanity checks.
>  >  checkMounted $USBDEV
>  >  if [ -n "$format" ];then
>  > +  checkLVMActive $USBDEV
>  >   # checks for a valid filesystem
>  >   if [ -n "$efi" ];then
>  >     createGPTLayout $USBDEV
>  > --
>  > 1.6.6.1
>  >
>  > --
>  > livecd mailing list
>  > livecd(a)lists.fedoraproject.org
>  > https://admin.fedoraproject.org/mailman/listinfo/livecd
>  --
>  livecd mailing list
>  livecd(a)lists.fedoraproject.org
>  https://admin.fedoraproject.org/mailman/listinfo/livecd

[Alan Pevec]

On Thu, Apr 22, 2010 at 11:00 PM, Sebastian Dziallas
<sdz(a)fedoraproject.org&gt; wrote:
>  I'm going to make some noise here as this is already the second
> patch
>  that has hit the list which needs a reaction 
Please review outstanding patches and give ack. I'd like to see
patches are publicly reviewed before being committed.
We could swap patch reviews: David has this and other (edit-livecd)
patches outstanding, and I just posted one.
Frederick, please review David's patches and he can review yours.
Besides Frederick's patch are there any other outstanding?

Thanks,
Alan

[Mads Kiilerich]

On 04/23/2010 01:05 PM, Alan Pevec wrote:
>  Frederick, please review David's patches and he can review
> yours.
>  Besides Frederick's patch are there any other outstanding? 
I proposed and outlined a patch on 
http://lists.fedoraproject.org/pipermail/livecd/2010-March/005714.html , 
but no authoritative conclusion was reached. Some "maintainer" feedback 
could be nice.

/Mads

[Alan Pevec]

On Thu, Apr 22, 2010 at 7:03 PM, David Huff <dhuff(a)redhat.com&gt; wrote:
>  +       for vg in $(pvs -o vg_name --noheadings $dev
> $dev${dev_delim}[0-9]* 2>/dev/null|sort -u); do
>  +               if [ $(lvs -o lv_attr --noheadings $vg |awk {'print $1;'} |cut -c
> 5) = "a" ]; then 
That check breaks with no or more than 1 LV but anyway, we should bail
out even on non active VG, it's not nice to reformat the disk under
LVM feet, so just first check is enough:

if [ "$(pvs -o vg_name --noheadings $dev $dev[0-9]* 2>/dev/null)" ]; then
	echo "Device, $dev, contains a volume group and cannot be formated!"
...

[PATCH livecd-tools] livecd-iso-to-disk --format fails with existing LV on USB

[Alan Pevec]

From: Alan Pevec <apevec(a)redhat.com&gt;

Resolves: rhbz#583658

Signed-off-by: Alan Pevec <apevec(a)redhat.com&gt;
---
 tools/livecd-iso-to-disk.sh |   13 +++++++++++++
 1 files changed, 13 insertions(+), 0 deletions(-)

diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh
index 3a4726d..f4d0855 100755
--- a/tools/livecd-iso-to-disk.sh
+++ b/tools/livecd-iso-to-disk.sh
@@ -138,6 +138,18 @@ checkPartActive() {
     fi
 }
 
+checkLVM() {
+    dev=$1
+
+    if [ -x /sbin/pvs -a \
+	"$(/sbin/pvs -o vg_name --noheadings $dev* 2>/dev/null)" ]; then
+	echo "Device, $dev, contains a volume group and cannot be formated!"
+	echo "You can remove the volume group using vgremove."
+	exitclean
+    fi
+    return 0
+}
+
 createGPTLayout() {
     dev=$1
     getdisk $dev
@@ -407,6 +419,7 @@ fi
 # do some basic sanity checks.  
 checkMounted $USBDEV
 if [ -n "$format" ];then
+  checkLVM $USBDEV
   # checks for a valid filesystem
   if [ -n "$efi" ];then
     createGPTLayout $USBDEV
-- 
1.6.0.6

[PATCH livecd-tools] livecd-iso-to-disk --format fails with existing LV on USB

[Frederick Grose]

On Fri, May 7, 2010 at 12:28 PM, <apevec(a)gmail.com&gt; wrote:

>  From: Alan Pevec <apevec(a)redhat.com&gt;
>
>  Resolves: rhbz#583658
>
>  Signed-off-by: Alan Pevec <apevec(a)redhat.com&gt;
>  ---
>   tools/livecd-iso-to-disk.sh |   13 +++++++++++++
>   1 files changed, 13 insertions(+), 0 deletions(-)
>
>  diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh
>  index 3a4726d..f4d0855 100755
>  --- a/tools/livecd-iso-to-disk.sh
>  +++ b/tools/livecd-iso-to-disk.sh
>  @@ -138,6 +138,18 @@ checkPartActive() {
>      fi
>   }
>
>  +checkLVM() {
>  +    dev=$1
>  +
>  +    if [ -x /sbin/pvs -a \
>  +       "$(/sbin/pvs -o vg_name --noheadings $dev* 2>/dev/null)" ]; then
>  +       echo "Device, $dev, contains a volume group and cannot be
>  formated!"
>  +       echo "You can remove the volume group using vgremove."
>  +       exitclean
>  +    fi
>  +    return 0
>  +}
>  +
>   createGPTLayout() {
>      dev=$1
>      getdisk $dev
>  @@ -407,6 +419,7 @@ fi
>   # do some basic sanity checks.
>   checkMounted $USBDEV
>   if [ -n "$format" ];then
>  +  checkLVM $USBDEV
>    # checks for a valid filesystem
>    if [ -n "$efi" ];then
>      createGPTLayout $USBDEV
>  --
>  1.6.0.6 

Redirecting standard error leads to silent failure when the lvm2 tools are
not installed.

I've suggested leaving out that redirection so as to leave a hint for the
user in that situation.

           --Fred

[PATCH livecd-tools] livecd-iso-to-disk --format fails with existing LV on USB

[Frederick Grose]

On Fri, May 7, 2010 at 1:04 PM, Frederick Grose <fgrose(a)gmail.com&gt; wrote:

>  On Fri, May 7, 2010 at 12:28 PM, <apevec(a)gmail.com&gt; wrote:
>
> > From: Alan Pevec <apevec(a)redhat.com&gt;
> >
> > Resolves: rhbz#583658
> >
> > Signed-off-by: Alan Pevec <apevec(a)redhat.com&gt;
> > ---
> >  tools/livecd-iso-to-disk.sh |   13 +++++++++++++
> >  1 files changed, 13 insertions(+), 0 deletions(-)
> >
> > diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh
> > index 3a4726d..f4d0855 100755
> > --- a/tools/livecd-iso-to-disk.sh
> > +++ b/tools/livecd-iso-to-disk.sh
> > @@ -138,6 +138,18 @@ checkPartActive() {
> >     fi
> >  }
> >
> > +checkLVM() {
> > +    dev=$1
> > +
> > +    if [ -x /sbin/pvs -a \
> > +       "$(/sbin/pvs -o vg_name --noheadings $dev* 2>/dev/null)" ];
> then
> > +       echo "Device, $dev, contains a volume group and cannot be
> > formated!"
> > +       echo "You can remove the volume group using vgremove."
> > +       exitclean
> > +    fi
> > +    return 0
> > +}
> > +
> >  createGPTLayout() {
> >     dev=$1
> >     getdisk $dev
> > @@ -407,6 +419,7 @@ fi
> >  # do some basic sanity checks.
> >  checkMounted $USBDEV
> >  if [ -n "$format" ];then
> > +  checkLVM $USBDEV
> >   # checks for a valid filesystem
> >   if [ -n "$efi" ];then
> >     createGPTLayout $USBDEV
> > --
> > 1.6.0.6
>
>
>  Redirecting standard error leads to silent failure when the lvm2 tools are
>  not installed.
>
>  I've suggested leaving out that redirection so as to leave a hint for the
>  user in that situation.
>
>             --Fred
>  
Sorry, I see the additional test now.
         --Fred

[PATCH livecd-tools] livecd-iso-to-disk --format fails with existing LV on USB

[David Huff]

On 05/07/2010 01:05 PM, Frederick Grose wrote:
>  On Fri, May 7, 2010 at 1:04 PM, Frederick
> Grose<fgrose(a)gmail.com>  wrote:
>
> > On Fri, May 7, 2010 at 12:28 PM,<apevec(a)gmail.com>  wrote:
> >
> >> From: Alan Pevec<apevec(a)redhat.com>
> >>
> >> Resolves: rhbz#583658
> >>
> >> Signed-off-by: Alan Pevec<apevec(a)redhat.com>
> >> ---
> >>   tools/livecd-iso-to-disk.sh |   13 +++++++++++++
> >>   1 files changed, 13 insertions(+), 0 deletions(-)
> >>
> >> diff --git a/tools/livecd-iso-to-disk.sh b/tools/livecd-iso-to-disk.sh
> >> index 3a4726d..f4d0855 100755
> >> --- a/tools/livecd-iso-to-disk.sh
> >> +++ b/tools/livecd-iso-to-disk.sh
> >> @@ -138,6 +138,18 @@ checkPartActive() {
> >>      fi
> >>   }
> >>
> >> +checkLVM() {
> >> +    dev=$1
> >> +
> >> +    if [ -x /sbin/pvs -a \
> >> +       "$(/sbin/pvs -o vg_name --noheadings $dev* 2>/dev/null)" ];
> then
> >> +       echo "Device, $dev, contains a volume group and cannot be
> >> formated!"
> >> +       echo "You can remove the volume group using vgremove."
> >> +       exitclean
> >> +    fi
> >> +    return 0
> >> +}
> >> +
> >>   createGPTLayout() {
> >>      dev=$1
> >>      getdisk $dev
> >> @@ -407,6 +419,7 @@ fi
> >>   # do some basic sanity checks.
> >>   checkMounted $USBDEV
> >>   if [ -n "$format" ];then
> >> +  checkLVM $USBDEV
> >>    # checks for a valid filesystem
> >>    if [ -n "$efi" ];then
> >>      createGPTLayout $USBDEV
> >> --
> >> 1.6.0.6
> >
> >
> > Redirecting standard error leads to silent failure when the lvm2 tools are
> > not installed.
> >
> > I've suggested leaving out that redirection so as to leave a hint for the
> > user in that situation.
> >
> >             --Fred
> >
>
>  Sorry, I see the additional test now.
>            --Fred 
committed!!

[PATCH livecd-tools] livecd-iso-to-disk --format fails with existing LV on USB

Frederick Grose wrote:
>      +    if [ -x /sbin/pvs -a \
>      +       "$(/sbin/pvs -o vg_name --noheadings $dev* 2>/dev/null)" ]; then
...
>  Redirecting standard error leads to silent failure when the lvm2
> tools 
>  are not installed. 
That's why there's a check for executable first.
Anyway, if lvm2 tools aren't installed, there won't be LVM started, so no
problem.

>  I've suggested leaving out that redirection so as to leave a hint
> for 
>  the user in that situation. 
That would be just noise and wouldn't help the user, IMHO it's better not to get
into error situation by checking for executable.

Thanks,
Alan