NetVault-Celerra虚拟带库NDMP备份与恢复
花了一下午才配置成功,不容易啊。
来自存储人生:《http://www.storageonline.com.cn/ ... e-with-celerra-vtl/》
虚拟磁带库(VTL)
VTL把磁盘虚拟成带库,在磁盘介质上模拟出机械臂、磁带机和磁带来进行工作。VTL 技术是数据备份领域的里程碑,具备多项优势,如性能高、故障率低、可靠性高、成本投入低以及运营成本低等,受到各行业用户的青睐。同时,VTL 也可以和现有磁带库集成,提高数据保护的整体安全性和性能,降低数据保护成本。
VTL 对于其它系统来说,与真实的、自动化的物理磁带库/机设备一样,区别仅在于数据是存储在磁盘上。并且,采用VTL 之后,大多数数据中心都明显感觉到,现在的数据备份进程更为快捷。于是,越来越多的用户开始淘汰掉原来的机械磁带库,采用虚拟磁带库作为主要的备份设备。
VTL与磁盘到磁盘直接备份相比,VTL具有两大优点:管理方便、性能良好。 在经过一段时间的发展后,VTL在性能方面的进步已经能够满足用户应用的需要,甚至可以说,最高端的VTL性能已经超越了最高端磁带库的性能,用户完全可以根据自己的需要进行选择。除此之外,VTL基本不存在机械故障的问题,这让维护也更加简单。同时,随着整合技术的进步,让VTL也具备了更加突出的技术优势。如在与SATA磁盘技术的结合后,可以提供大容量的解决方案;加入iSCSI后,VTL可以实现异地备份,对数据量小的应用环境更具有吸引力;与重复数据删除技术结合后,还可以大幅度降低实际存储的数据容量。
VTL市场现有厂商状况
目前市场上的的VTL产品主要由两大类厂商提供:
VTL软件厂商:
FalconStor:EMC、IBM、SUN、同有、众志合达、浪潮等多家OEM
Sepaton:自有产品、HP OEM
Alacritus:已被NetApp收购
Diligent:HDS OEM,Overland OEM、SUN OEM
NearTek:EMC开始收购Neartek公司的VTL资产。
VTL硬件厂商:
国际:EMC 、NetApp、HP、IBM 、Quantum、Sepaton、SUN、StorgeTek、Spectra Logic、Mirage、Maxxan、Copan Systems。
国内:同有、众志合达、浪潮等
来自《存储部落》
网络数据管理协议(NDMP)
网络数据管理协议NDMP是一种开放式标准,用于集中控制企业级数据管理。NDMP体系结构可使备份应用程序厂商在Network Appliance Filer和其他网络连接服务器上控制本机备份和恢复设备。NDMP最初由NetApp与Legato Systems(现在的EMC Networker)公司联合开发设计。
详情请参考《网络数据管理协议NDMP原理详解》和《Configuring NDMP Backups on Celerra》
Celerra NDMP backups to Disk
简介:
The Celerra® Network Server’s virtual tape library unit (VTLU) for NDMP allows a Celerra Network Server to back up data directly to disk-based storage by means of the NDMP protocol. The EMC® VTLU feature expands the functionality of the industry-standard NDMP protocol to include compatibility with disk-based storage, allowing you to manage physical tape library units (TLUs) and VTLUs from the same NDMP-compatible backup software.
A VTLU resides on a Data Mover, and communicates with the backup software in the same manner as a physical TLU. The backup software, for example, EMC NetWorker®, cannot distinguish between a VTLU and a physical TLU. The E-Lab™ Interoperability Navigator provides information on qualified backup software vendors.
限制:
Checkpoint file systems cannot be imported into a VTLU.
To implement an NDMP backup-to-disk solution, the disk storage must be attached to the Celerra Network Server and not to the backup server running
the backup software.
The server_archive and server_mt commands are not compatible with a VTLU.
The server_devconfig command does not discover VTLUs; it discovers only physical devices.
Use dedicated file systems for VTLU storage.
Auto extend should be set to “No”.
逻辑图:
步骤:
Celerra上新建VTL;
Celerra配置ndmp账号;
NetVault添加NDMP server;
NetVault添加Device;
NetVault测试NDMP Backup to Celerra VTL;
NetVault测试NDMP Restore from Celerra VTL;
测试环境:EMC Celerra NS-G8(6.0.41-3), Bakbone NetVault Backup(Version 8.22 Build 29 Release R2009AUG06-GENUS)
1. Celerra上新建VTL
1.1 Celerra新建文件系统
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
nasadmin@G8_CS0 ~]$ nas_fs -name jonathan_vtlu -c size=10G pool=clar_r5_performance
id = 1408
name = jonathan_vtlu
acl = 0
in_use = False
type = uxfs
worm = off
volume = v769
pool = clar_r5_performance
member_of = root_avm_fs_group_3
rw_servers=
ro_servers=
rw_vdms =
ro_vdms =
auto_ext = no,virtual_provision=no
deduplication = unavailable
stor_devs = CKM00090900315-000B,CKM00090900315-0009
disks = d29,d26
1.2 Celerra挂载该文件系统
1
2
[nasadmin@G8_CS0 ~]$ server_mount server_2 jonathan_vtlu /jonathan_vtlu
server_2 : done
1.3 Celerra新建VTL
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
[nasadmin@G8_CS0 ~]$ server_vtlu server_2 -tlu -new -robot -vendor jonathan
-product jonathan_p -revision 2.0
server_2 :
id = 7
slots = 32
import/export slots = 8
robot vendor = jonathan
robot product = jonathan_p
robot revision = 2.0
robot serial number = 1a9OYQH3k5
robot device name = c0t1l4
drives = 4
drive vendor = EMC
drive product = svtlDrive
drive revision = 1.01
列举刚建的VTL:
1
2
3
4
[nasadmin@G8_CS0 ~]$ server_vtlu server_2 -tlu -l
server_2 :
id vendor product revision serial_number device_name
7 jonathan jonathan_p 2.0 1a9OYQH3k5 c0t1l4
1.4 Celerra给VTL指定存储空间
这里如果不指定tapesize,默认为50GB。如果不指定tapes,默认会创建最多的tape。
1
2
3
[nasadmin@G8_CS0 ~]$ server_vtlu server_2 -storage -new jonathan_vtlu -tlu 7
-tapesize 1G -tapes 5 -barcodeprefix jonathantape
server_2 : done
查看VTL存储空间:
1
2
3
4
[nasadmin@G8_CS0 ~]$ server_vtlu server_2 -storage -l 7
server_2 :
tlu_id filesystem barcode_prefix
7 jonathan_vtlu jonathantape
查看VTL tape:
1
2
3
4
5
6
7
8
[[nasadmin@G8_CS0 ~]$ server_vtlu server_2 -tape -l 7
server_2 :
barcode filesystem capacity(GB) location source_slot
jonathantape00jonathan_vtlu 1 slot:0
jonathantape00jonathan_vtlu 1 slot:1
jonathantape00jonathan_vtlu 1 slot:2
jonathantape00jonathan_vtlu 1 slot:3
jonathantape00jonathan_vtlu 1 slot:4
查看VTL drive:
1
2
3
4
5
6
7
[nasadmin@G8_CS0 ~]$ server_vtlu server_2 -drive -l 7
server_2 :
drive_id device_name serial_number status tape_barcode
0 c0t1l5 ZYr8bvb200 empty
1 c0t1l6 fvQgcvb200 empty
2 c0t1l7 v1fpdvb200 empty
3 c0t1l8 LxDxevb200 empty
2. Celerra配置ndmp账号
这里用户名一定是ndmp,且采用MD5加密,否则restore时会报错。
1
[nasadmin@G8_CS0 ~]$ /nas/sbin/server_user server_2 -add -md5 -passwd ndmp
3. NetVault添加NDMP server
4. NetVault添加Device
这里要指定Library的tape为EMC Celerra NDMP2D:
这里要给每一个drive指定唯一的drive bay号:
保存成功:
5. NetVault测试NDMP Backup to Celerra VTL
这里有几个参数可以设,详细设置参见参考文档:
运行Backup的job,没有问题:
6. NetVault测试NDMP Restore from Celerra VTL
这是文件系统roger_fs_02现在的文件:
1
2
3
4
5
6
7
8
[root@G8_CS0 roger_fs_02]# ll -h –full-time
total 24K
-rw-rw-r– 1 park service 0 2011-05-28 12:12:22.000001000 +0800 file1
drwxr-xr-x 3 root bin 80 2011-05-20 18:22:52.000017000 +0800 homedir
-rw-r–r– 1 root root 0 2011-06-01 18:56:39.000006000 +0800 jonathan
drwxr-xr-x 2 root root 8.0K 2011-05-07 15:03:58.000002000 +0800 lost+found
drwxrwxrwx 4 32769 32768 1.0K 2011-05-28 23:21:59.000000000 +0800 test
drwxr-xr-x 2 32769 32768 80 2011-05-07 15:08:12.000008000 +0800 test11
删掉其中一个文件:
1
2
3
4
5
6
7
8
9
[root@G8_CS0 roger_fs_02]# rm jonathan
rm: remove regular empty file `jonathan’? y
[root@G8_CS0 roger_fs_02]# ll -h –full-time
total 24K
-rw-rw-r– 1 park service 0 2011-05-28 12:12:22.000001000 +0800 file1
drwxr-xr-x 3 root bin 80 2011-05-20 18:22:52.000017000 +0800 homedir
drwxr-xr-x 2 root root 8.0K 2011-05-07 15:03:58.000002000 +0800 lost+found
drwxrwxrwx 4 32769 32768 1.0K 2011-05-28 23:21:59.000000000 +0800 test
drwxr-xr-x 2 32769 32768 80 2011-05-07 15:08:12.000008000 +0800 test11
我们来尝试用刚才Backup的Saveset来恢复文件系统:
恢复成功:
查看恢复后的文件系统,发现删掉的文件被成功恢复:
1
2
3
4
5
6
7
8
root@G8_CS0 roger_fs_02]# ll -h –full-time
total 24K
-rw-rw-r– 1 park service 0 2011-05-28 12:12:22.000001000 +0800 file1
drwxr-xr-x 3 root bin 80 2011-05-20 18:22:52.000017000 +0800 homedir
-rw-r–r– 1 root root 0 2011-06-01 18:56:39.000001000 +0800 jonathan
drwxr-xr-x 2 root root 8.0K 2011-05-07 15:03:58.000002000 +0800 lost+found
drwxrwxrwx 4 32769 32768 1.0K 2011-05-28 23:21:59.000000000 +0800 test
drwxr-xr-x 2 32769 32768 80 2011-05-07 15:08:12.000008000 +0800 test11
参考文档:
http://docs.bakbone.com/docs/Net ... for_EMC_Celerra.pdf
Configuring NDMP Backups on Celerra
http://www.bakbone.com/docs/NDMP_Backups_with_Celerra.pdf
Configuring NDMP Backups to Disk on Celerre
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论