Moving clients from one networker server to another

i prefer to add storage nodes to scale, since it is not quite so easy to move backups between 2 backup servers (and thus different datazones). Why is your backup server overloaded? try to shift load from that backup server to a (new) storage node.

You can also opt not to do any backups to the backup server(expect it's own ofcourse); this could drastically reduce the load on your backup server by relegating it role just to schedule & track metadata of all backups (so not to fuzz around with terabytes of actual data on the backup server.)

René

Hello Rene,

We already have 4 storage nodes, and we are not taking backups to the backup server itself, only to the storage nodes.

But with more and more clients, and thus more and more savesets, the media database, jobsdb, ... are all growing, and it seems this is starting to give problems.

We already lowered the retention of the jobsdbd to 3 days, and a maximum size of 60MB, but we see that it has a size of 200-300MB just to keep info for 3 days.

Also several mminfo's are taking an very long time to finish, and while running such mminfo querry's, the NMC has trouble loading it's info.

Hello Bobby,

That was exactly what I was looking for.

Too bad there is not an easy way to move the current backups to the new server. We are working with adv_file devices on datadomain.

So I guess we'll have to move all the backups of those clients to a specific device, and mount that device on the new server, and do a complete scanner of that device.

We also got that tuning guide from our support. And applied all the settings and recomandations.

Kind regards

Could you please let me know what process requires to be done if I add new storage nodes? How do I make sure those are being used? How is the migration done?

If this is a new client/storage node ...

  - Install the NW client and storage node software

  - Prepare its /nsr/res/servers file appropriately

      If it has changed, restart the NW client service/daemon

  - Define the client in your data zone

  - Define at least one remote device at that client

  - Add the appropriate license(s)

If this is an existing client ...

  - Install/promote the NW client by installing the storage node software

  - Verify the /nsr/res/servers file

      If it has changed, restart the NW client service/daemon

  - Define at least one remote device at that client

  - Add the appropriate license(s)

Finally, do not forget

  - to define additional pools (if required)

  - label new media to that pool

  - Adjust the client resources (storage node list) for the clients which shall use the storage node appropriately

Of course you can :

  - do not forget that the media db does 'only' store save set and media information, but no device information

  - tapes will not be affected at all. That's why you can use them in any compatible device

  - disk drives will only be affected if you label/recycle the media. This is what you must not do.

Excuse me, but let me ask directly: What do you want at all?

1. question:   How to add another storage node?

2. question:   Can I move devices to the SN?

3. question:   Can DDs be moved at all?

    Sure they can - but how do you want to move them at all?

      - Within the same or to a different data zone?

      - The whole DD or a single device on the DD?

      - Device and/or media as well?

      - Keeping existing data or not?

    If you just want to create a new DD device for the new SN you may of course also do that.

    If it shall receive data for the same pool you may also stop writing to the existing DD device/volume as follows:

      - UNMOUNT the current DD volume

      - set the volume to 'read only'

      - MOUNT - DO NOT LABEL - it again

It will lead to an endless discussion if you just mention every detail in a separate question instead letting us know the scenario.

If you have DDs, there is no reason to have so many devices for sure.  Anyway, back to you questions:

1. remove device (do not remove volume data) and add this device again to storage node (without labeling of course) - then just mount it and data (volume) is now available via storage node
2. You need storage licenses for each storage node IF you use legacy licensing (I say this as you mentioned 362 devices and if your server was storage node and you never had extra storage node licenses to bump up parallelism streams, it means that you could only support 64 streams at max which may be the reason why you saw hangs in the first place).
3. Just change storage node field of the client
4. Nope

Also when you say if its DD there is no reason to have so many devices. Do you mean its enough to create say 1 device for 1 pool, and the rest will be taken care by the data domain?

This worked. Thanks a lot! Hrvoje Crvelin 

I use 4-5 devices per pool (and have 2 pools only) which is enough to ingest 50k sessions per day.