Networker integration with CloudBoost

Networker and CloudBoostLets discuss how CloudBoost with ECS in the backend can be integrated with a Networker.

CloudBoost and ECS integration posts is here: part 1, part 2

The typical case for a CloudBoost is a backup cloning. I’ll configure such config first.

The initial thing we have to do is to create Storage Node for CloudBoost.Networker and CloudBoost

All we need is to specify CB FQDN.Networker and CloudBoost

CB Storage Node is created.Networker and CloudBoost

The next step is a Device creation. Open a New Device Wizard with a right click.4

Device type is AFTD.5

Storage Node = CB FQDN. Keep “Browse storage node or network path” selected.6

/mnt/magfs/base path has to be selected.7

Lets name the device as device0.8

Keep “Label and Mount device after creation” checked. Pool Type = Backup Clone. Pool = Default Clone.9

Check the configuration.15

CB device is created, labeled and mounted to a Default Clone Pool.11

Corresponding folder structure is created.

maginatics@cb:/mnt/magfs/base$ cd /mnt/magfs/base/
maginatics@cb:/mnt/magfs/base$ ls
00  04  08  12  16  20  24  28  32  36  40  44  48  52  56  60  64  68  72  76  80  84  88  92  96  active
01  05  09  13  17  21  25  29  33  37  41  45  49  53  57  61  65  69  73  77  81  85  89  93  97  volhdr
02  06  10  14  18  22  26  30  34  38  42  46  50  54  58  62  66  70  74  78  82  86  90  94  98
03  07  11  15  19  23  27  31  35  39  43  47  51  55  59  63  67  71  75  79  83  87  91  95  99
maginatics@cb:/mnt/magfs/base$ cat volhdr
@ַV@ַPw??VTD`V@ַZ=?@ַV@ַPw??Vnw.vipr.local.001,
volume poolDefaultClone

Next create a device for a backup. For my test I’d like to make it local AFTD on the Networker server Storage Node.12

In my config /opt/nsr is a separate big file system. That is why I created “media” directory there.

[root@nw linux_x86_64]# mkdir /opt/nsr/media

Device name is device1.13

Keep “Label and Mount device after creation” checked. Pool Type = Backup. Pool = Default.14

Check the device1 config.15

Now both backup device1 is created.16

Corresponding structure is created as well.

[root@nw linux_x86_64]# ls /opt/nsr/media/
00  04  08  12  16  20  24  28  32  36  40  44  48  52  56  60  64  68  72  76  80  84  88  92  96  active
01  05  09  13  17  21  25  29  33  37  41  45  49  53  57  61  65  69  73  77  81  85  89  93  97  volhdr
02  06  10  14  18  22  26  30  34  38  42  46  50  54  58  62  66  70  74  78  82  86  90  94  98
03  07  11  15  19  23  27  31  35  39  43  47  51  55  59  63  67  71  75  79  83  87  91  95  99

Navigate to Configuration > Clients.17

Just for testing purposes modify the client with All Save Set. I’d like to backup my Networker Linux distribution located in the /opt/nsr/backup directory.Double click on the client. Change Save Set to /opt/nsr/backup.18

Check if the Save Set was changed.19

Navigate to Configuration > Groups.20

Double click on the Default Group. Check that Clone Pool = Default Clone. Clones shouldn’t be checked in my tests because I’m going to start cloning manually.21

Manually start the Default Group backup with a right click on it.22

Navigate to Monitoring. Check a backup progress. Backup transfer rate is 20MB/s.23

Double click on the Group and check if a backup was successful with no Failed Save Sets.25

Navigate to Configuration > Clones.26

Create new Clone with CBclone name. Storage Node to WRITE save sets = our CloudBoost Storage node. Write clone data to pool = Default Clone.27

Clone is created.28

Right click on the CBclone to start it.29

Navigate to Monitoring > Clones > Clones tab.30

You can see how CB capacity is consumed with time. Check the /mnt/magfs filesystem.

maginatics@cb:/mnt/magfs/base$ df -k
Filesystem                                1K-blocks    Used Available Use% Mounted on
/dev/sda1                                   9480420 4214100          4761696  47% /
none 4       0                4   0% /sys/fs/cgroup
udev 8213832       4          8213828   1% /dev
tmpfs 1643428     572          1642856   1% /run
none 5120       0             5120   0% /run/lock
none 8217136       0          8217136   0% /run/shm
none 102400       0           102400   0% /run/user
/dev/mapper/TrollGroup-TrollVolume         41149760 1183664         39949712   3% /lvm
//cloud.emc.com/cloud              9007199254740928       0 9007199254740928   0% /mnt/magfs

maginatics@cb:/mnt/magfs/base$ df -k
Filesystem                                1K-blocks    Used Available Use% Mounted on
/dev/sda1                                   9480420 4268388          4707408  48% /
none                                              4       0                4   0% /sys/fs/cgroup
udev 8213832       4          8213828   1% /dev
tmpfs 1643428     572          1642856   1% /run
none 5120       0             5120   0% /run/lock
none 8217136       0          8217136   0% /run/shm
none 102400       0           102400 0% /run/user
/dev/mapper/TrollGroup-TrollVolume         41149760 1189396         39943980   3% /lvm
//cloud.emc.com/cloud              9007199254740928  516160 9007199254224768   1% /mnt/magfs

Double click on the CBclone to check if the cloning was successful.31

504MB were written on CB. Dedup ration is 1.0.32

Lets make another full backup of the same data. Always Full level is configured in the Default Group properties.33

Start and check the backup.34

Next start a cloning of the same data.35

Interesting, it is very fast in comparison with the previous cloning. The reason is dedup.

Checking CB dashboard we can confirm that dedup ratio now is 2.0. Blobstore size stays the same but Share size was doubled.36

OK Cloning works as expected.

The next question if we can make a backup directly to a CB? Lets check that. First we have to re-label our CB device. Navigate to Devices.37

Instead of Default Clone, CB AFTD device should belong to the Default (backup) Pool.38

Also we need to unmounts device1 which also belongs to Default Pool.39

device1 is unmounted.40

Navigate to Configuration > Clients and edit the client we already tested for a backup.

In the Globals (2 of 2) tab we should insert our CB FQDN in the list of Storage Nodes. It should be in front of nsrserverhost.41

Navigate to Monitoring and start the Default Group.42

Data are backed up directly to the CB with 932KB/s transfer rate (ECS is located somewhere in the Internet).43

Interesting that dedup ration is 0.5 now. The reason is: after re-labeling Networker swiped the Share. But Blobstore was not cleared.44

Actual CB capacity confirms actually CB capacity was cleared after re-label and the actual capacity corresponds to the size of the last backup.

maginatics@cb:/mnt/magfs/base$ df -k
Filesystem                                1K-blocks    Used Available Use% Mounted on
/dev/sda1                                   9480420 4288692          4687104  48% /
none 4       0                4   0% /sys/fs/cgroup
udev 8213832       4          8213828   1% /dev
tmpfs                                       1643428     572          1642856   1% /run
none 5120       0             5120   0% /run/lock
none 8217136       0          8217136   0% /run/shm
none 102400       0           102400   0% /run/user
/dev/mapper/TrollGroup-TrollVolume         41149760 1192160         39941216   3% /lvm
//cloud.emc.com/cloud              9007199254740928  515968 9007199254224960   1% /mnt/magfs

After a while (about 1 hour) Blobstore info was fixed. I think that is done by the CB garbage collection.45

Yoo-hoo! I’m done with Networker -> CloudBoost -> ECS testing…

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: