Previously the sync for dump_fs is problematic, it always return success according to man 2 sync. So it cannot detect the error of the dump target is full and not all of vmcore data been written back the disk, which will leave the vmcore imcomplete and report misleading log as "saving vmcore complete".
In this patch, we will use "sync -f vmcore" instead, which will return error if syncfs on the dump target fails. In this way, vmcore sync related failures, such as autoextend of lvm2 thinpool fails, can be detected and handled properly.
Signed-off-by: Tao Liu ltao@redhat.com --- dracut-kdump.sh | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/dracut-kdump.sh b/dracut-kdump.sh index f4456a1..343e2b0 100755 --- a/dracut-kdump.sh +++ b/dracut-kdump.sh @@ -175,8 +175,14 @@ dump_fs() _dump_exitcode=$? if [ $_dump_exitcode -eq 0 ]; then mv "$_dump_fs_path/vmcore-incomplete" "$_dump_fs_path/vmcore" - sync - dinfo "saving vmcore complete" + sync -f "$_dump_fs_path/vmcore" + _sync_exitcode=$? + if [ $_sync_exitcode -eq 0 ]; then + dinfo "saving vmcore complete" + else + derror "sync vmcore failed, exitcode:$_sync_exitcode" + return 1 + fi else derror "saving vmcore failed, exitcode:$_dump_exitcode" return 1
On Fri, May 27, 2022 at 02:40:40PM +0800, Tao Liu wrote:
Previously the sync for dump_fs is problematic, it always return success according to man 2 sync. So it cannot detect the error of the dump target is full and not all of vmcore data been written back the disk, which will leave the vmcore imcomplete and report misleading log as "saving vmcore complete".
In this patch, we will use "sync -f vmcore" instead, which will return error if syncfs on the dump target fails. In this way, vmcore sync related failures, such as autoextend of lvm2 thinpool fails, can be detected and handled properly.
Signed-off-by: Tao Liu ltao@redhat.com
dracut-kdump.sh | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/dracut-kdump.sh b/dracut-kdump.sh index f4456a1..343e2b0 100755 --- a/dracut-kdump.sh +++ b/dracut-kdump.sh @@ -175,8 +175,14 @@ dump_fs() _dump_exitcode=$? if [ $_dump_exitcode -eq 0 ]; then mv "$_dump_fs_path/vmcore-incomplete" "$_dump_fs_path/vmcore"
I think it is better to call "syncfs" on "vmcore-incomeplete" and not "vmcore". If this file can be synced properly, then one should move it to vmcore.
In fact, that's the whole point of vmcore-incomplete. If dump could not be saved successfuly, we will see vmcore-incomplete instead of vmcore.
BTW, where is the upstream git repo for this code now a days?
Thanks Vivek
syncdinfo "saving vmcore complete"
sync -f "$_dump_fs_path/vmcore"_sync_exitcode=$?if [ $_sync_exitcode -eq 0 ]; thendinfo "saving vmcore complete"elsederror "sync vmcore failed, exitcode:$_sync_exitcode"return 1fi-- 2.33.1
Hi Vivek,
On Fri, May 27, 2022 at 8:56 PM Vivek Goyal vgoyal@redhat.com wrote:
On Fri, May 27, 2022 at 02:40:40PM +0800, Tao Liu wrote:
Previously the sync for dump_fs is problematic, it always return success according to man 2 sync. So it cannot detect the error of the dump target is full and not all of vmcore data been written back the disk, which will leave the vmcore imcomplete and report misleading log as "saving vmcore complete".
In this patch, we will use "sync -f vmcore" instead, which will return error if syncfs on the dump target fails. In this way, vmcore sync related failures, such as autoextend of lvm2 thinpool fails, can be detected and handled properly.
Signed-off-by: Tao Liu ltao@redhat.com
dracut-kdump.sh | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/dracut-kdump.sh b/dracut-kdump.sh index f4456a1..343e2b0 100755 --- a/dracut-kdump.sh +++ b/dracut-kdump.sh @@ -175,8 +175,14 @@ dump_fs() _dump_exitcode=$? if [ $_dump_exitcode -eq 0 ]; then mv "$_dump_fs_path/vmcore-incomplete" "$_dump_fs_path/vmcore"
I think it is better to call "syncfs" on "vmcore-incomeplete" and not "vmcore". If this file can be synced properly, then one should move it to vmcore.
In fact, that's the whole point of vmcore-incomplete. If dump could not be saved successfuly, we will see vmcore-incomplete instead of vmcore.
Agreed, thanks for pointing this out, I will syncfs against vmcore-incomplete.
BTW, where is the upstream git repo for this code now a days?
For fedora kexec-tools rpm git repo:
git clone https://src.fedoraproject.org/rpms/kexec-tools.git
Thanks, Tao Liu
Thanks Vivek
syncdinfo "saving vmcore complete"
sync -f "$_dump_fs_path/vmcore"_sync_exitcode=$?if [ $_sync_exitcode -eq 0 ]; thendinfo "saving vmcore complete"elsederror "sync vmcore failed, exitcode:$_sync_exitcode"return 1fi else derror "saving vmcore failed, exitcode:$_dump_exitcode" return 1-- 2.33.1
-
Tao Liu -
Vivek Goyal