Lost Management IP

[fr.mueller]

Hi,

i am playing around with the backup and restore feature in UCS.

After restoring a system from sratch with the full-state backup, the Management IP for the Fabric B Interconnect is missing. 

It seems to be, that the ip is not restored. Is that a known issue?

Are there any other point to keep in mind for a restore from scratch? Are there other points missing? Or do i have to restore the other backup files as well?

Thanks for sharing

Frank

[Walter Dey]

Hallo Frank

Did you erase the FI, and then

Enter the configuration method. (console/gui) ? console
Enter the setup mode; setup newly or restore from backup. (setup/restore) ? restore

----------------------------------------------------------------------------------------------------------------

What I usually do, and this is also a cool procedure to clone other UCS domains

- Do a "All configuration" backup; this is a xml, not like the "Full State"; it can be modified or only partially used for restore; important "Preserve Identities" must be set

- Erase config and/or go to a brand new UCS domain;

- do initial basic setup, IP, mask Gateway, admin pw

- login to UCSM and do a restore,

 

[fr.mueller]

Hi,

yes, i do a complete erase configuration.

Today, i do it again. With a service profile connected to a server.

After restoring the FI´s from scratch, the management IP is lost. OK thats not a big issue.

But the biggest issue is, that the nics and hba attached to my service profile are not working. They are error disabled and stuff like that.

I have to first disassociate the service profile and unbind from the template. After that, i re-assigned the template and the server boot up. Bevor, the server start up stops at the bios.

What the hell is that? What, if i have to restore my FI´s with maybe 50 hosts?

Thats not a backup and restore solution. Or am i missing something?

Frank

[Walter Dey]

Frank

Don't do a full backup, but a backup "all configuration" with preserve identities, like a mention above.

Follow the procedure above; I have done this multiple times, it works; even with SP and boot from SAN !

[fr.mueller]

Hi,

today, after doing the erase configuration thing, i restore the fi only with my all config backup.

But there are some configs missing.

1. My certificate is (Key Ring Settings) missing

2. The Service Profile provided is unassociated. Configuration error: server unavailable. There are not enough ressources overall. I have to associate the profile again with "change service profile association" (Maybe i am doing something wrong here.)

3. Domain authentication also not working

Any thoughts?

Frank

[Walter Dey]

Frank

questions

- did you do a "backup all configuration" AND "preserve identites" set ?

- after erase configuration: did you do a basic setup, ip address, mask, gwy, pw for admin

- the start UCSM

- do a import configuration; action = replace !

[fr.mueller]

Hi,

here are the steps.

1. Erase configuration on FI-A and FI-B

2. Initial Config, setup IP, mask gw etc.

3. Add second FI

4. Wait for completing the setup

5. Import configuration from all-configuration with preserved identities.

6. Wait for reboot to complete.

7. wan´t to login with my AD credentials. Not possible

8. My own created certificate is missing

9. Service Profile error on my blade server

Reimport my certiticate

Unbind from Template for the blade and bind it again.

I am currently checking why the AD integration is not working.

UPDATE:

The Password for the LDAP Provider User is not stored in the backup. Therefor i have to add it manually!

Frank

[Walter Dey]

In my test, I was only using local accounts.

Regarding LDAP, I would open a TAC case, seems to be bug !

What exactly is the error message of the SP (your point 9)

I don't understand why a unbind / rebind from the template is necessary ?

[fr.mueller]

The error message is

There are not enough ressources overall

Frank

[Walter Dey]

OK, but often, you have additional information in the message, which resources are missing ? e.g. vnic, mac,.......