Unsolved
This post is more than 5 years old
91 Posts
1
3443
Help me please with using second Storage Node
Can you help me with EMC Networker and use second Storage node. I have used Networker 8.2.3.3 on server (include server and storage node,console) with local own AFTD storage. I want to make additional storage on new server with storage node. I install client on new server and chose storage node role. Then I in my Networker Server in console add storage node and create new device. Now i have 2 device, several client, several group where i choose "Use networker server for backup" and don't use pools (blank). Here my question: How i can configure my groups\clients, that I can make backup on different time to different server with storage node : Example: Client_1 included in Group1, in started backup at 10:00PM to Networker server and at 10:00AM to storage node (Server2).
Thanks!
ble1
14.3K Posts
0
October 18th, 2016 07:00
Only way to do this is to have both storage nodes (server1 and server2) in storage node list, different pools on each and have one group assigned to one pool and second one to second. However, why complicate it? If you want security in case that aftd storage fails on one side so you have it at least on other one, then rather use cloning between devices on server1 and server2.
Mandri
91 Posts
0
October 19th, 2016 05:00
Hi Hrvoje Crvelin!
How I know, I can't assign one group to one pool, I have to assign one client record to one pool. Then create second same client end assign another pool. Is it correct way to resolve my problem ?
bingo.1
2.4K Posts
0
October 19th, 2016 06:00
Please familiarize with pools.
In fact you can assign to certain pool ...
- a group (all save sets which have been created by running the group)
- a client
- a save set
- a level
- the combination of 2 or more of these selection criteria.
Defining a pool is nothing else but filtering a certain save stream.
Don't forget that the more criteria you select the more you narrow the filter. This often leads to mis-pooling.
The other issue Hrvoje already mentioned is that this limits your failover capabilities. In other words:
If you tell NW to send a specific save stream to a certain pool, other data will just not be accepted.
So in case the remote SN fails for whatever reason NW will not have any choice to backup if you only have
one device/media prepared at this SN. And if the SN fails the data cannot be backed up by the other pool at
the server. NW will just wait until the problem will be solved.
However, if you have online media of the same pool at both nodes, Use the client's storage node list to define the preferred SN sequence. In case the first one fails the other can continue.
Mandri
91 Posts
0
October 19th, 2016 07:00
Thanks you!
And a last question, what about this article https://community.emc.com/docs/DOC-31696 ?
Must I use a list (list of storage node) on each client to have opportunity to use filter case : example: client+ pools + group = backup direction (needed Storage Node) ?
bingo.1
2.4K Posts
0
October 19th, 2016 07:00
no - please verify that with your existing client resources. It is an optional list which will be followed top down.
Mandri
91 Posts
0
October 21st, 2016 02:00
Hi All.
Can anyone show me example steps, how configure backup a folder to two different storage node ?
I had tried, but all time receive same error: Unable to setup direct save with server server.local: no matching devices for save of client `server.local; check storage nodes, devices or pools.
I don't understand how use 2 storage node where each one in different polls ? (data save in the one SP, index can't save in default SP - have error).
I see the only way: create 2 clients end configure to use each different pool. Is it correct way ? (a don't know, how use groups to choose different storage node)
bingo.1
2.4K Posts
0
October 21st, 2016 03:00
The most obvious issue is that you specify the client as selection criteria for the pool. This of course is o.k. ... for the data save sets. However, the client INDEX and the BOOTSTRAP save sets are belonging to the NWSERVER's client - they just live/belong there so this is the logical consequence.
What is the outcome?
- If you have a pool 'Client_ABC' with selection criteria "Client_ABC", this pool will only accept data from "Client_ABC".
Exactly as you specified.
- The index however belongs to "Client_NWSERVER" and is rejected by pool "Client_ABC" with the above criteria.
How can you change the behavior?
- You could add "Client_NWSERVER" to the selection criteria 'client' list. This would work ... but not for all backups
because the bootstrap still must be backed up by a locally attached device at the NWSERVER.
- Define another pool for index and bootstrap backups with media only available at NWSERVER.
Selection criteria would only be this 'save set' list:
index:*
bootstrap
BTW - this is best practice as all metadata will be stored in a separate pool.
It does not make sense to ask for such help as we do not know your configuration. Your should provide that for comments.
Mandri
91 Posts
0
October 21st, 2016 04:00
Bingo, thanks for your replay, but I didn't understand how if have 1 client, 2 storage pool, 1-2 groups make process for backup client to 2 different way (storage pool_1 and storage pool_2 at different time ).
thanks .
bingo.1
2.4K Posts
0
October 22nd, 2016 02:00
The brief solution is:
- 2 groups
- each for one client resource
- 3 pools
- 2 data pools (1 for each SN). Use only the group name as selection criteria
- 1 index pool
Is that sufficient info? - I doubt that.
If you want to know why may I suggest that you download the following doc
ftp.legato.com/pub/NetWorker/German/tdocs/archive/nw_cookbook_jan_2009.pdf
The screenshots are outdated as they will show the old admin GUI but the general info is still valid.
Then go to page 241 and .... do not forget to read ;-) .
Mandri
91 Posts
0
October 24th, 2016 06:00
Hi bingo!
Do you know, how client link to pool ? I read a lot of instructions, but all this instructions says about Client resource (without pool) and pools with groups (without Client).
If we don't choose a pool for client, clients saveset will save to default pool ( server storage node).
I understand, so pools has properties: Data Source group, clients, SaveSet, device and level....
P.S. I make a lot of tests and identified: without name of second storage node in client properties. Why no one don't say about that, only say - read documentation........
bingo.1
2.4K Posts
0
October 24th, 2016 23:00
In general, I urge you not to select the pool inside the client resource.
This is a property which has been added later - I am pretty sure that it has not been used very often.
"If we don't choose a pool for client, clients saveset will save to default pool ( server storage node)."
Not necessarily - only if there is no other pool which will fit. See the doc for examples.
"I understand, so pools has properties: Data Source group, clients, SaveSet, device and level...."
Obviously. As a pool exists to filter data it makes sense to provide filter criteria. Correct?
"P.S. I make a lot of tests and identified: without name of second storage node in client properties. Why no one don't say about that, only say - read documentation......"
As you obviously never attended a training, you must simply read the fundamental information to understand how pools work in general. Once you got the principle (which an Admin Guide will not tell you), you will see that understanding pools is not so difficult but it is pretty easy to 'mis-pool' if you do not know how they work.