[lustre-discuss] Backup software for Lustre
Dilger, Andreas
andreas.dilger at intel.com
Sun Mar 19 14:11:52 PDT 2017
The use of openat() could be problematic since this precludes storing the xattr before the file is opened. That said, I don't see anywhere in your strace log that (f)setxattr() is called to restore the xattrs, for either the regular files or directories, even after the file is opened or written?
Does the RHEL tar have a whitelist of xattrs to be restored? The fact that there are Lustre xattrs after the restore appears to just be normal behavior for creating a file, not anything related to tar restoring xattrs.
Cheers, Andreas
On Mar 19, 2017, at 10:45, Brett Lee <brettlee.lustre at gmail.com<mailto:brettlee.lustre at gmail.com>> wrote:
Sure, happy to help. I did not see mknod+setxattr in the strace output. Included is a trimmed version of the strace output, along with a few more bits of information. Thanks!
# cat /proc/fs/lustre/version
lustre: 2.7.19.8
# cat /etc/redhat-release
CentOS Linux release 7.3.1611 (Core)
# uname -r
3.10.0-514.2.2.el7_lustre.x86_64
# rpm -qa|grep tar
tar-1.26-31.el7.x86_64
# sha1sum `which tar` `which gtar`
ea17ec98894212b2e2285eb2dd99aad76185ea7d /usr/bin/tar
ea17ec98894212b2e2285eb2dd99aad76185ea7d /usr/bin/gtar
Striping was set on the four directories before creating the files.
mkdir -p /scratch/1; lfs setstripe -c 1 --stripe-size 128K /scratch/1; lfs getstripe /scratch/1
mkdir -p /scratch/2; lfs setstripe -c 2 --stripe-size 256K /scratch/2; lfs getstripe /scratch/2
mkdir -p /scratch/3; lfs setstripe -c 3 --stripe-size 768K /scratch/3; lfs getstripe /scratch/3
mkdir -p /scratch/4; lfs setstripe -c 4 --stripe-size 1M /scratch/4; lfs getstripe /scratch/4
After tar, all files and directories all had the default Lustre striping.
# tar ztvf /scratch.tgz
drwxr-xr-x root/root 0 2017-03-19 10:54 scratch/
drwxr-xr-x root/root 0 2017-03-19 10:57 scratch/4/
-rw-r--r-- root/root 4194304 2017-03-19 10:57 scratch/4/4.dd
drwxr-xr-x root/root 0 2017-03-19 10:57 scratch/3/
-rw-r--r-- root/root 4194304 2017-03-19 10:57 scratch/3/3.dd
drwxr-xr-x root/root 0 2017-03-19 10:57 scratch/1/
-rw-r--r-- root/root 4194304 2017-03-19 10:57 scratch/1/1.dd
drwxr-xr-x root/root 0 2017-03-19 10:57 scratch/2/
-rw-r--r-- root/root 4194304 2017-03-19 10:57 scratch/2/2.dd
# strace tar zxvf /scratch.tgz > strace.out 2>&1
execve("/usr/bin/tar", ["tar", "zxvf", "/scratch.tgz"], [/* 22 vars */]) = 0
...
(-cut - loading libraries)
...
fstat(1, {st_mode=S_IFREG|0644, st_size=10187, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f4a63d9f000
write(1, "scratch/\n", 9scratch/
) = 9
mkdirat(AT_FDCWD, "scratch", 0700) = -1 EEXIST (File exists)
newfstatat(AT_FDCWD, "scratch", {st_mode=S_IFDIR|0755, st_size=4096, ...}, AT_SYMLINK_NOFOLLOW) = 0
write(1, "scratch/4/\n", 11scratch/4/
) = 11
mkdirat(AT_FDCWD, "scratch/4", 0700) = 0
write(1, "scratch/4/4.dd\n", 15scratch/4/4.dd
) = 15
openat(AT_FDCWD, "scratch/4/4.dd",
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
O_WRONLY|O_CREAT|O_EXCL|O_NOCTTY|O_NONBLOCK|O_CLOEXEC, 0600) = 4
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 5632) = 5632
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 5632) = 5632
...
(-cut)
...
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 512) = 512
dup2(4, 4) = 4
fstat(4, {st_mode=S_IFREG|0600, st_size=4194304, ...}) = 0
utimensat(4, NULL, {{1489935825, 0}, {1489935444, 0}}, 0) = 0
fchown(4, 0, 0) = 0
fchmod(4, 0644) = 0
close(4) = 0
write(1, "scratch/3/\n", 11scratch/3/
) = 11
newfstatat(AT_FDCWD, "scratch/4", {st_mode=S_IFDIR|0700, st_size=4096, ...}, AT_SYMLINK_NOFOLLOW) = 0
utimensat(AT_FDCWD, "scratch/4", {{1489935825, 0}, {1489935444, 0}}, AT_SYMLINK_NOFOLLOW) = 0
fchownat(AT_FDCWD, "scratch/4", 0, 0, AT_SYMLINK_NOFOLLOW) = 0
fchmodat(AT_FDCWD, "scratch/4", 0755) = 0
mkdirat(AT_FDCWD, "scratch/3", 0700) = 0
write(1, "scratch/3/3.dd\n", 15scratch/3/3.dd
) = 15
openat(AT_FDCWD, "scratch/3/3.dd", O_WRONLY|O_CREAT|O_EXCL|O_NOCTTY|O_NONBLOCK|O_CLOEXEC, 0600) = 4
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 6656) = 6656
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
...
(-cut - pick up with last file...)
...
d(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
--- SIGCHLD {si_signo=SIGCHLD, si_code=CLD_EXITED, si_pid=2476, si_status=0, si_utime=7, si_stime=0} ---
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 10240) = 10240
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 7680) = 7680
dup2(4, 4) = 4
fstat(4, {st_mode=S_IFREG|0600, st_size=4194304, ...}) = 0
utimensat(4, NULL, {{1489935825, 0}, {1489935432, 0}}, 0) = 0
fchown(4, 0, 0) = 0
fchmod(4, 0644) = 0
close(4) = 0
clock_gettime(CLOCK_REALTIME, {1489935825, 628399394}) = 0
clock_gettime(CLOCK_REALTIME, {1489935825, 628414336}) = 0
close(3) = 0
wait4(2476, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0, NULL) = 2476
newfstatat(AT_FDCWD, "scratch/2", {st_mode=S_IFDIR|0700, st_size=4096, ...}, AT_SYMLINK_NOFOLLOW) = 0
utimensat(AT_FDCWD, "scratch/2", {{1489935825, 0}, {1489935432, 0}}, AT_SYMLINK_NOFOLLOW) = 0
fchownat(AT_FDCWD, "scratch/2", 0, 0, AT_SYMLINK_NOFOLLOW) = 0
fchmodat(AT_FDCWD, "scratch/2", 0755) = 0
newfstatat(AT_FDCWD, "scratch", {st_mode=S_IFDIR|0755, st_size=4096, ...}, 0) = 0
utimensat(AT_FDCWD, "scratch", {{1489934977, 0}, {1489935261, 0}}, 0) = 0
fchownat(AT_FDCWD, "scratch", 0, 0, 0) = 0
close(1) = 0
munmap(0x7f4a63d9f000, 4096) = 0
close(2) = 0
exit_group(0) = ?
+++ exited with 0 +++
Brett
--
Protect Yourself Against Cybercrime
PDS Software Solutions LLC
https://www.TrustPDS.com<https://www.trustpds.com/>
On Sun, Mar 19, 2017 at 7:39 AM, Dilger, Andreas <andreas.dilger at intel.com<mailto:andreas.dilger at intel.com>> wrote:
I ran a test locally with RHEL 6.8 and the included tar 1.26 using strace, and tar is properly using mknod+setxattr to restore the "lov" xattr, and the stripe count and stripe size to be preserved.
The OST index is not preserved with the xattr restore, since that may cause imbalance if the files were backed up in a different filesystem (e.g. one with fewer OSTs). The MDS will balance OST allocation as needed for the current OST usage.
Could you please run your tar on RHEL 7 with strace to see if it is doing this correctly?
Cheers, Andreas
On Mar 18, 2017, at 21:51, Brett Lee <brettlee.lustre at gmail.com<mailto:brettlee.lustre at gmail.com>> wrote:
Hi Andreas, I expected that to be the case, but found out it was not. Instead, the restore restores everything - unless directed otherwise.
Backup == cmd + add xattrs.
Restore == cmd + exclude xattrs.
Brett
--
Protect Yourself Against Cybercrime
PDS Software Solutions LLC
https://www.TrustPDS.com
On Mar 18, 2017 9:28 PM, "Dilger, Andreas" <andreas.dilger at intel.com<mailto:andreas.dilger at intel.com>> wrote:
Do you need to specify --xattrs (or similar) during the restore phase as well?
Cheers, Andreas
On Mar 17, 2017, at 15:12, Brett Lee <brettlee.lustre at gmail.com<mailto:brettlee.lustre at gmail.com>> wrote:
Hi. In what I thought was a valid test, I was unable to confirm that a backup and restore retained the layouts. Perhaps my expectation or process was incorrect? The process was:
1. Create 4 files, each with different stripe sizes and stripe counts (verified with getstripe).
2. Back up the files using tar-1.26-31.el7.x86_64.
3. Recreate a file system and restore the files.
Backup command: tar --xattrs -zcvf /scratch.tgz /scratch
Restore command: tar zxvf /scratch.tgz
After restoration, getstripe showed that each file had the default stripe count (1) and stripe size (1MB).
FWIW: After restoring, getfattr produced the same result for each file:
# getfattr -d -m - -R <file>
lustre.lov=0s0AvRCwEAAAAdAAAAAAAAAAAEAAACAAAAAAAQAAEAAAAFAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
trusted.link=0s3/HqEQEAAAAuAAAAAAAAAAAAAAAAAAAAABYAAAACAAAEAAAAAAUAAAAAMS5kZA==
trusted.lma=0sAAAAAAAAAAAABAAAAgAAAB0AAAAAAAAA
trusted.lov=0s0AvRCwEAAAAdAAAAAAAAAAAEAAACAAAAAAAQAAEAAAAFAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=
Brett
--
Protect Yourself Against Cybercrime
PDS Software Solutions LLC
https://www.TrustPDS.com<https://www.trustpds.com/>
On Wed, Mar 15, 2017 at 5:03 AM, Dilger, Andreas <andreas.dilger at intel.com<mailto:andreas.dilger at intel.com>> wrote:
I believe Zmanda is already using GNU tar (or RHEL tar) for the actual backup storage? I that case it should already work, since we fixed tar long ago to backup and restore xattrs in a way that preserves Lustre layouts.
Cheers, Andreas
On Mar 14, 2017, at 15:47, Brett Lee <brettlee.lustre at gmail.com<mailto:brettlee.lustre at gmail.com>> wrote:
Thanks for the details, Andreas!
Maybe OpenSFS can fund Zmanda so that their backup software can include the Lustre metadata... :)
Brett
--
Protect Yourself Against Cybercrime
PDS Software Solutions LLC
https://www.TrustPDS.com<https://www.trustpds.com/>
On Tue, Mar 14, 2017 at 3:13 PM, Dilger, Andreas <andreas.dilger at intel.com<mailto:andreas.dilger at intel.com>> wrote:
To reply to this old thread, there are two different kinds of Lustre backup solutions:
- file level backups that traverse the client POSIX filesystem, for which any number of
commercial solutions exist. Making these solutions "capable of saving Lustre metadata"
boils down to two simple things - save the "lustre.lov" xattr during backup (at a minimum,
other xattrs also should be backed up), and then using mknod(2) + setxattr() to restore
the "lustre.lov" xattr before opening the file and restoring the data.
- device level backups (e.g. "dd" for ldiskfs, and "zfs send/recv" for ZFS).
Using the file level backups allows backup/restore of subsets of the filesystem, since many
HPC sites have Lustre filesystems that are too large to backup completely. I typically do
not recommend to use device-level backups for the OSTs, unless doing an OST hardware migration,
and even then it is probably less disruptive to do Lustre-level file migration off the OST
before swapping it out.
Whether file level backups are used or not, I would recommend sites always make periodic
device level backups of the MDT(s). The amount of space needed for an MDT backup is small
compared to the rest of the filesystem (e.g. a few TB at most), and can avoid the need for
a full filesystem restore (e.g. multi-PB of data, if a full backup exists at all) even
though all the data is still available on the OSTs.
The MDT device-level backup can use relatively slow SATA drives, since they will mostly be
used for linear writes (or occasionally linear reads for restore), so a few multi-TB SATA III
drives is sufficient for storing a rotating set of MDT device backups. At 150MB/s for even
a single SATA drive, this is about 2h/TB, which is reasonable to do once a week (or more often
for smaller MDTs).
While using an LVM snapshot of the ldiskfs MDT for the backup source is desirable for consistency
reasons, having even an MDT backup from a mounted and in-use MDT is better than nothing at
all when a problem is hit, since e2fsck can repair the in-use inconsistencies fairly easily,
and Lustre can deal with inconsistencies between the MDT and OST reasonably (at most returning
an -ENOENT error to the client for files that were deleted).
Cheers, Andreas
On Feb 7, 2017, at 12:32, Andrew Holway <andrew.holway at gmail.com<mailto:andrew.holway at gmail.com>> wrote:
>
> Would it be difficult to suspend IO and snapshot all the nodes (assuming ZFS). Could you be sure that your MDS and OSS are synchronised?
>
> On 7 February 2017 at 19:52, Mike Selway <mselway at cray.com<mailto:mselway at cray.com>> wrote:
>> Hello Brett,
>>
>> Actually, looking for someone who uses a commercialized approach (that retains user metadata and Lustre extended metadata) and not specifically the manual approaches of Chapter 17.
>>
>> Thanks!
>> Mike
>>
>> Mike Selway | Sr. Tiered Storage Architect | Cray Inc.
>> Work +1-301-332-4116<tel:%2B1-301-332-4116> | mselway at cray.com<mailto:mselway at cray.com>
>> 146 Castlemaine Ct, Castle Rock, CO 80104 | www.cray.com<http://www.cray.com>
>>
>>
>>> From: Brett Lee [mailto:brettlee.lustre at gmail.com<mailto:brettlee.lustre at gmail.com>]
>>> Sent: Monday, February 06, 2017 11:45 AM
>>> To: Mike Selway <mselway at cray.com<mailto:mselway at cray.com>>
>>> Cc: lustre-discuss at lists.lustre.org<mailto:lustre-discuss at lists.lustre.org>
>>> Subject: Re: [lustre-discuss] Backup software for Lustre
>>>
>>> Hey Mike,
>>>
>>> "Chapter 17" and
>>> http://www.intel.com/content/www/us/en/lustre/backup-and-restore-training.html
>>>
>>> both contain methods to backup & restore the entire Lustre file system.
>>>
>>> Are you looking for a solution that backs up only the (user) data files and their associated metadata (e.g. xattrs)?
>>>
>>> Brett
>>> --
>>> Protect Yourself From Cybercrime
>>> PDS Software Solutions LLC
>>> https://www.TrustPDS.com
>>>
>>>> On Mon, Feb 6, 2017 at 11:12 AM, Mike Selway <mselway at cray.com<mailto:mselway at cray.com>> wrote:
>>>>
>>>> Hello,
>>>> Anyone aware of and/or using a Backup software package to protect their LFS environment (not referring to the tools/scripts suggested in Chapter 17).
>>>>
>>>> Regards,
>>>> Mike
Cheers, Andreas
--
Andreas Dilger
Lustre Principal Architect
Intel Corporation
_______________________________________________
lustre-discuss mailing list
lustre-discuss at lists.lustre.org<mailto:lustre-discuss at lists.lustre.org>
http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20170319/42b51977/attachment-0001.htm>
More information about the lustre-discuss
mailing list