1 Rookie
•
5 Posts
0
1528
February 11th, 2022 07:00
Rename EMC Networker (Rebuild without bootstraps)
My predecessor created the networker server using the FQDN as its name. I now have to rename the domain which entails renaming the networker server. I have administrated networker for quite some time but have never used processes like scanner as i have never had the need. I would like your opinion as to my process to accomplish the manual renaming of my server. My server is currently live and I plan to record as much information from it as possible so any tips on what i should record would also be appreciated. Please look at my process below and let me know what you think and thank you for your time.
- Perform boot strap backup and record appropriate information to recover.
- Record all client ID’s minus the networker server itself.
- Note all retention information.
- Remove Networker and rename folders to OLD_EMC Networker.
- Install Networker all default locations as was the old server.
- Create the clients using the old client ID’s
- Make sure all the retention information is accurate to what was listed before.
- Initiate scanner process to repopulate the databases.
- Verify all databases reflect the appropriate information.
No Events found!
bingo.1
2.4K Posts
0
February 15th, 2022 01:00
It is a shame that nobody at Dell/ EMC has public tools to easily support such migration, especially after 30 years in the market. Maybe Professional Services know better.
Actually I do not know what will happen as I never looked at the retention dates. But even if the old values will not be used, you can still change the retention date (per save set) with the nsrmm command. Script this process and you are done.
bingo.1
2.4K Posts
0
February 11th, 2022 08:00
Which NW version do you use at all?
Nuckinfutz52
1 Rookie
•
5 Posts
0
February 11th, 2022 12:00
19.5.0.5
bingo.1
2.4K Posts
0
February 12th, 2022 03:00
It is not a secret, but - depending on your configuration - it can become a challenge. Especially if you cannot use the same directory. For example, on Windows, if NW has been installed under the default pathname, I strongly recommend that you move NW off the C: drive.
The other potential issue I see is the renaming of the NW server. This means that a lot of resources and entries in the media db need to be adjusted. This will be an obstacle. Please get in touch with Dell/EMC support how to overcome it.
Whatever you do - test it extensively before you use the hot system. Install some VMs where you can easily verify the behavior. And remember, NW will run in eval mode for 90 days, so you have enough time to do that. And you can test special scripts so that some steps can run much faster later. Of course they will avoid you from making manual mistakes.
This is how I would proceed (keeping the same hostname):
1. On the new system:
- Install NW (into the same directory)
- Identify the hostid and
- request a new license file/auth codes
2. On the old system:
- make sure you run a bootstrap at the last backup
This allows you to save the scanner processes. Please see footnote.
- run mminfo -aX > file
- run mminfo -avot > file
- run nsrls > file
- copy the files to another (the new) server
- disable all workflows (scripted ?)
- unmount all backup volumes (scripted ?)
- disable all devices (scripted ?)
- disable all storage nodes (scripted ?)
- stop NW
- if the NW volume resides on a VM, dismount this disk. Otherwise copy it to an external disk
3. On the new NW server:
- Mount the external disk or the virtual NW Disk
- Install NW into the same directory (if possible)
- stop NW if it has already (automatically) been started
- save and move /nsr/mm, /nsr/index & /nsr/res
- copy from old NW /nsr/mm, /nsr/index & /nsr/res
- rename /nsr/res/nsrladb
- restart NW and wait until everything is running (nsrwatch)
- Check the NW database by running and comparing the previous reports again
- Re-enable storage nodes & devices (scripted?)
- Re-mount the backup disk volumes (if available, scripted?)
- Enable only 1 workflow and start it
- If successful, re-enable all other workflows (scripted?)
- Start NMC and verify that it works
- Run new backups
- Run at least 1 restore to verify the general functionality
- Do not forget licensing in the end
As you can see, you should prepare a fine list with all necessary steps in the proper sequence.
Beware of scanning. Since NW 19.3, there is a bug where NW slows down extremely when running 'scanner -i' on DDBoost and AFTD devices which reside on a DD. May I suggest you read my thread "scanner -i" will take too long for DDBoost & DD-AFTD devices" in this forum.
Finally - test ... test ... test
Good luck
Nuckinfutz52
1 Rookie
•
5 Posts
0
February 14th, 2022 07:00
I have been told I cannot rename networker by Ell EMC support. What i am attempting to do is build a new server using the new name and using scanner recreate the indexes. I only have LTO tapes and a small amount of clients so im fairly sure this will work. I was hoping to find out if the retention dates will be retained when i perform the scanner process so my media expiration will be as it should.
Nuckinfutz52
1 Rookie
•
5 Posts
0
February 15th, 2022 07:00
That is kind of what i was thinking. Thank you.
clebeuva
1 Rookie
•
58 Posts
0
February 19th, 2022 06:00
FYI, we wanted Dell to perform a migration of our NetWorker server/NMC to a new server (different name) on a different domain. It was not possible to migrate as such. We had to recreate all policies/workflows on a new install.
We are using DataDomain, so they used the new volume move tool which was very efficient to export and then import all online saveset indexes and clients. As you say scanner would have been very slow.
bingo.1
2.4K Posts
0
February 21st, 2022 03:00
Well done. As you said, the volume move tool is pretty new stuff - afair it got introduced about 3 months ago.
BTW - in general, scanner is running pretty fast. Especially on DD based volumes. Unfortunately, with NW 19.3+ there is a bug which slows down the process on such media.
It is a shame that such a severe bug is not detected by QA (if it even exists). And it is even worse as nobody seems to take ownership to get that repaired.