23.07.2014 Views

Lustre 1.6 Operations Manual

Lustre 1.6 Operations Manual

Lustre 1.6 Operations Manual

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

How do I backup / restore a <strong>Lustre</strong> filesystem?<br />

Several types of <strong>Lustre</strong> backups are available.<br />

CLIENT FILE SYSTEM-LEVEL BACKUPS<br />

It is possible to back up <strong>Lustre</strong> filesystems from a client (or many clients in parallel<br />

working in different directories), via any number of user-level backup tools like tar,<br />

cpio, Amanda, and many enterprise-level backup tools. However, due to the very<br />

large size of most <strong>Lustre</strong> filesystems, full backups are not always possible. Doing<br />

backups of subsets of the filesystem (subdirectories, per user, incremental by date,<br />

etc.) using normal file backup tools is still recommended, as this is the easiest<br />

method from which to restore data.<br />

TARGET RAW DEVICE-LEVEL BACKUPS<br />

In some cases, it is desirable to do full device-level backups of an individual MDS or<br />

OST storage device for various reasons (before hardware replacement, maintenance<br />

or such). Doing full device-level backups ensures that all of the data is preserved in<br />

the original state and is the easiest method of doing a backup.<br />

If hardware replacement is the reason for the backup or if there is a spare storage<br />

device then it is possible to just do a raw copy of the MDS/OST from one block<br />

device to the other as long as the new device is at least as large as the original device<br />

using the command:<br />

dd if=/dev/{original} of=/dev/{new} bs=1M<br />

If hardware errors are causing read problems on the original device then using the<br />

command below allows as much data as possible to be read from the original device<br />

while skipping sections of the disk with errors:<br />

dd if=/dev/{original} of=/dev/{new} bs=4k conv=sync,noerror<br />

Even in the face of hardware errors, the ext3 filesystem is very robust and it may be<br />

possible to recover filesystem data after e2fsck is run on the new device.<br />

TARGET FILE SYSTEM-LEVEL BACKUPS<br />

In other cases, it is desirable to make a backup of just the file data in an MDS or OST<br />

filesystem instead of backing up the entire device (e.g., if the device is very large but<br />

has little data in it, if the configuration of the parameters of the ext3 filesystem need<br />

to be changed, to use less space for the backup, etc).<br />

In this case it is possible to mount the ext3 filesystem directly from the storage<br />

device and do a file-level backup. <strong>Lustre</strong> MUST BE STOPPED ON THAT NODE.<br />

To back up such a filesystem properly also requires that any extended attributes<br />

(EAs) stored in the filesystem be backed up, but unfortunately current backup tools<br />

do not properly save this data so an extra step is required.<br />

Appendix D <strong>Lustre</strong> Knowledge Base D-13

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!