If nofail or nobootwait option is used, systemd's local-fs.target won't wait for the mounting to complete, and kdump might start before the required mount point is ready and then fail.
The host might use nofail for reasons like the device may get unpluged, and if the device is not mounted and it is set as kdump target as the same time then kdump service won't start, we will never enter the capture kernel. By the time we have entered the capture kernel, the target device must exist and ready to use, or else kdump would fail anyway. So force remove nofail and nobootwait option.
Signed-off-by: Kairui Song kasong@redhat.com --- mkdumprd | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/mkdumprd b/mkdumprd index 3d89a80..078f988 100644 --- a/mkdumprd +++ b/mkdumprd @@ -104,6 +104,9 @@ to_mount() { _options=$(echo $_options | sed 's/\bnoauto\b//') #mount fs target as rw in 2nd kernel _options=$(echo $_options | sed 's/\bro\b/rw/') + # drop nofail or nobootwait + _options=$(echo $_options | sed 's/\bnofail\b//') + _options=$(echo $_options | sed 's/\bnobootwait\b//')
_mntopts="$_target $_fstype $_options" #for non-nfs _dev converting to use udev persistent name
Hi Kairui, On 08/09/18 at 01:27pm, Kairui Song wrote:
If nofail or nobootwait option is used, systemd's local-fs.target won't wait for the mounting to complete, and kdump might start before the required mount point is ready and then fail.
The host might use nofail for reasons like the device may get unpluged, and if the device is not mounted and it is set as kdump target as the same time then kdump service won't start, we will never enter the capture kernel. By the time we have entered the capture kernel, the target device must exist and ready to use, or else kdump would fail anyway. So force remove nofail and nobootwait option.
Signed-off-by: Kairui Song kasong@redhat.com
mkdumprd | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/mkdumprd b/mkdumprd index 3d89a80..078f988 100644 --- a/mkdumprd +++ b/mkdumprd @@ -104,6 +104,9 @@ to_mount() { _options=$(echo $_options | sed 's/\bnoauto\b//') #mount fs target as rw in 2nd kernel _options=$(echo $_options | sed 's/\bro\b/rw/')
# drop nofail or nobootwait
_options=$(echo $_options | sed 's/\bnofail\b//')
_options=$(echo $_options | sed 's/\bnobootwait\b//')
_mntopts="$_target $_fstype $_options" #for non-nfs _dev converting to use udev persistent name
For some reason we added "nofail" in /etc/sysconfig/kdump, see kdump.sysconfig.$(uname -m) in git tree. But maybe we do not need it anymore, can you check it, if not needed, remove from sysconfig as well in this patch?
Thanks Dave
Hi Dave,
That nofail used for rootflag, we don't always require root to be ready, like when dumping to a network a target or raw target. But when fstab is used, and we are dumping to an fs target, the target has to be ready before kdump.sh is called.
nofail flag was first introduced in 7c48f71 to avoid sysroot.mount failure blocks kdump script, and in that commit nofail was added to fstab entries too. But later c7ed468 removed the nofail in fstab due to the same problem this patch is trying to solve, kdump.sh got called before target is ready.
I think we could just keep the nofail for rootflags.
On Fri, Aug 10, 2018 at 10:13 AM Dave Young dyoung@redhat.com wrote:
Hi Kairui, On 08/09/18 at 01:27pm, Kairui Song wrote:
If nofail or nobootwait option is used, systemd's local-fs.target won't wait for the mounting to complete, and kdump might start before the required mount point is ready and then fail.
The host might use nofail for reasons like the device may get unpluged, and if the device is not mounted and it is set as kdump target as the
same
time then kdump service won't start, we will never enter the capture kernel. By the time we have entered the capture kernel, the target device must exist and ready to use, or else kdump would fail anyway. So force remove nofail and nobootwait option.
Signed-off-by: Kairui Song kasong@redhat.com
mkdumprd | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/mkdumprd b/mkdumprd index 3d89a80..078f988 100644 --- a/mkdumprd +++ b/mkdumprd @@ -104,6 +104,9 @@ to_mount() { _options=$(echo $_options | sed 's/\bnoauto\b//') #mount fs target as rw in 2nd kernel _options=$(echo $_options | sed 's/\bro\b/rw/')
# drop nofail or nobootwait
_options=$(echo $_options | sed 's/\bnofail\b//')
_options=$(echo $_options | sed 's/\bnobootwait\b//')
_mntopts="$_target $_fstype $_options" #for non-nfs _dev converting to use udev persistent name
For some reason we added "nofail" in /etc/sysconfig/kdump, see kdump.sysconfig.$(uname -m) in git tree. But maybe we do not need it anymore, can you check it, if not needed, remove from sysconfig as well in this patch?
Thanks Dave
On 08/10/18 at 11:08am, Kairui Song wrote:
Hi Dave,
That nofail used for rootflag, we don't always require root to be ready, like when dumping to a network a target or raw target. But when fstab is used, and we are dumping to an fs target, the target has to be ready before kdump.sh is called.
nofail flag was first introduced in 7c48f71 to avoid sysroot.mount failure blocks kdump script, and in that commit nofail was added to fstab entries too. But later c7ed468 removed the nofail in fstab due to the same problem this patch is trying to solve, kdump.sh got called before target is ready.
When nofail root flags was added we still can not remove root dependency, even if dumping to network dracut still mount sysroot. But now we only mount root when necessary, for example use root fs as a dump target. So in current situation if mounting root then root is a must so I think we can drop that flag.
I think we could just keep the nofail for rootflags.
On Fri, Aug 10, 2018 at 10:13 AM Dave Young dyoung@redhat.com wrote:
Hi Kairui, On 08/09/18 at 01:27pm, Kairui Song wrote:
If nofail or nobootwait option is used, systemd's local-fs.target won't wait for the mounting to complete, and kdump might start before the required mount point is ready and then fail.
The host might use nofail for reasons like the device may get unpluged, and if the device is not mounted and it is set as kdump target as the
same
time then kdump service won't start, we will never enter the capture kernel. By the time we have entered the capture kernel, the target device must exist and ready to use, or else kdump would fail anyway. So force remove nofail and nobootwait option.
Signed-off-by: Kairui Song kasong@redhat.com
mkdumprd | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/mkdumprd b/mkdumprd index 3d89a80..078f988 100644 --- a/mkdumprd +++ b/mkdumprd @@ -104,6 +104,9 @@ to_mount() { _options=$(echo $_options | sed 's/\bnoauto\b//') #mount fs target as rw in 2nd kernel _options=$(echo $_options | sed 's/\bro\b/rw/')
# drop nofail or nobootwait
_options=$(echo $_options | sed 's/\bnofail\b//')
_options=$(echo $_options | sed 's/\bnobootwait\b//')
_mntopts="$_target $_fstype $_options" #for non-nfs _dev converting to use udev persistent name
For some reason we added "nofail" in /etc/sysconfig/kdump, see kdump.sysconfig.$(uname -m) in git tree. But maybe we do not need it anymore, can you check it, if not needed, remove from sysconfig as well in this patch?
Thanks Dave
-- Best Regards, Kairui Song
Thanks Dave
Good point, will update the patch.
On Fri, Aug 10, 2018 at 12:02 PM Dave Young dyoung@redhat.com wrote:
On 08/10/18 at 11:08am, Kairui Song wrote:
Hi Dave,
That nofail used for rootflag, we don't always require root to be ready, like when dumping to a network a target or raw target. But when fstab is used, and we are dumping to an fs target, the target
has
to be ready before kdump.sh is called.
nofail flag was first introduced in 7c48f71 to avoid sysroot.mount
failure
blocks kdump script, and in that commit nofail was added to fstab entries too. But later c7ed468 removed the nofail in fstab due to the same
problem
this patch is trying to solve, kdump.sh got called before target is
ready.
When nofail root flags was added we still can not remove root dependency, even if dumping to network dracut still mount sysroot. But now we only mount root when necessary, for example use root fs as a dump target. So in current situation if mounting root then root is a must so I think we can drop that flag.
I think we could just keep the nofail for rootflags.
On Fri, Aug 10, 2018 at 10:13 AM Dave Young dyoung@redhat.com wrote:
Hi Kairui, On 08/09/18 at 01:27pm, Kairui Song wrote:
If nofail or nobootwait option is used, systemd's local-fs.target
won't
wait for the mounting to complete, and kdump might start before the required mount point is ready and then fail.
The host might use nofail for reasons like the device may get
unpluged,
and if the device is not mounted and it is set as kdump target as the
same
time then kdump service won't start, we will never enter the capture kernel. By the time we have entered the capture kernel, the target
device
must exist and ready to use, or else kdump would fail anyway. So
force
remove nofail and nobootwait option.
Signed-off-by: Kairui Song kasong@redhat.com
mkdumprd | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/mkdumprd b/mkdumprd index 3d89a80..078f988 100644 --- a/mkdumprd +++ b/mkdumprd @@ -104,6 +104,9 @@ to_mount() { _options=$(echo $_options | sed 's/\bnoauto\b//') #mount fs target as rw in 2nd kernel _options=$(echo $_options | sed 's/\bro\b/rw/')
# drop nofail or nobootwait
_options=$(echo $_options | sed 's/\bnofail\b//')
_options=$(echo $_options | sed 's/\bnobootwait\b//')
_mntopts="$_target $_fstype $_options" #for non-nfs _dev converting to use udev persistent name
For some reason we added "nofail" in /etc/sysconfig/kdump, see kdump.sysconfig.$(uname -m) in git tree. But maybe we do not need it anymore, can you check it, if not needed, remove from sysconfig as
well
in this patch?
Thanks Dave
-- Best Regards, Kairui Song
Thanks Dave
-
Dave Young -
Kairui Song