Hi, i have a doubt about mongodb backup using lvm snapshot.

In the manual
document(http://docs.mongodb.org/manual/administration/backups/), in the Backup
With Journaling section it says that if you enable journal you can backup
your database in following steps:
1.Create Snapshot

lvcreate --size 100M --snapshot --name mdb-snap01 /dev/vg0/mongodb

2.Archive Snapshots

umount /dev/vg0/mdb-snap01

dd if=/dev/vg0/mdb-snap01 | tar -czf mdb-snap01.tar.gz

3.Restore Snapshot

lvcreate --size 1G --name mdb-new vg0
tar -xzf mdb-snap01.tar.gz | dd of=/dev/vg0/mdb-new
mount /dev/vg0/mdb-new /srv/mongodb

4.Remote Backup Storage

umount /dev/vg0/mdb-snap01
dd if=/dev/vg0/mdb-snap01 | ssh usern...@example.com tar -czf
/opt/backup/mdb-snap01.tar.gz
lvcreate --size 1G --name mdb-new vg0
ssh usern...@example.com tar -xzf /opt/backup/mdb-snap01.tar.gz | dd of=
/dev/vg0/mdb-new
mount /dev/vg0/mdb-new /srv/mongodb

there  is not lock fsync operations necessary.

just as (http://www.mongodb.org/display/DOCS/Backups)said:As long as all
files are snapshotted at the same point in time, you don't need to
lock+fsync

but when taking lvm snapshot, everything are writen to the Journal log ,but
the data is stealing in the file system buffer or the OS buffer is that ok
for we to recover
our data to a consistent point .(especially when the sln No.s of  Journal
log and Filesystem snapshot are not consistent)

*looking forward for your answer.*
*
*
thank you very much .

Hi Jinxuan Hou --

As I understand the question, then by default, the 'journal commit
interval'  is 100 millisecond (unless you have overridden the 'journal
commit interval' value or if you issue getLastError command).  As such,
there may be ~100 millisecond window during which data durability is as
risk.  Also, if you are using Amazon EBS, refer to the section in the
manual.

Hope this answers your question.

Regards,

Kay