cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
161
Views
0
Helpful
4
Replies
Highlighted
Beginner

Service Profile Association Fails - "server unavailable There are not enough resources overall"

I have a UCS Mini running firmware 3.2(3k) and recently purchased two UCS-B200-M5 blades that are identical in configuration. One blade arrived first, and it was installed when the chassis was running firmware 3.1. I upgraded to 3.2(3k) as this was required to use the B200-M5.

 

Post firmware upgrade I was able to apply a service profile successfully and use the blade without any errors. A few days later we received the other identical blade and installed it physically in the server, but it will not allow a service profile to be associated with it. It gives the error "server unavailable There are not enough resources overall", without any specifics on what resource it is having an issue with. 

I thought it may be a firmware issue, but it will not allow me to install firmware to the blade, it gives the error "there is nothing to update please check the configuration". The firmware the blade is currently running is 4.0.2a, which I understand is supported as long as it is the same or later than the infrastructure firmware.

 

Any assistance or suggestions would be appreciated.

 

Thanks,


Derek

1 ACCEPTED SOLUTION

Accepted Solutions
Cisco Employee

Re: Service Profile Association Fails - "server unavailable There are not enough resources overall"

Greetings.

Please check the cross version support section in the ucsm release notes. https://www.cisco.com/c/en/us/td/docs/unified_computing/ucs/release/notes/CiscoUCSManager-RN-4-0.html#concept_rlp_4h2_zdb

You can run a newer UCSM, and older blade/host firmware version, but not the other way around, which is your situation.

If you are running 4.0x blade firmware, then you will need ucsm 4.0x UCSM.

 

Did the newer blade fully discover?

I would try creating a service profile with only minimal objects (i.e just 1 vnic), and see what objects cause the SP to fail.

Make sure you have enough available identities in various pools (UUID, MAC, WWPN, IPs) as that can cause a similar message.

Also, make sure you aren't using server pools, as I've seen similar issue when a service profile is referencing a server pool, that a specific physical server is not a member of.

 

There are various logs in UCSM that will ID the issue pretty quick but not via the context of a public forum.

 

Kirk..

4 REPLIES 4
Cisco Employee

Re: Service Profile Association Fails - "server unavailable There are not enough resources overall"

Greetings.

Please check the cross version support section in the ucsm release notes. https://www.cisco.com/c/en/us/td/docs/unified_computing/ucs/release/notes/CiscoUCSManager-RN-4-0.html#concept_rlp_4h2_zdb

You can run a newer UCSM, and older blade/host firmware version, but not the other way around, which is your situation.

If you are running 4.0x blade firmware, then you will need ucsm 4.0x UCSM.

 

Did the newer blade fully discover?

I would try creating a service profile with only minimal objects (i.e just 1 vnic), and see what objects cause the SP to fail.

Make sure you have enough available identities in various pools (UUID, MAC, WWPN, IPs) as that can cause a similar message.

Also, make sure you aren't using server pools, as I've seen similar issue when a service profile is referencing a server pool, that a specific physical server is not a member of.

 

There are various logs in UCSM that will ID the issue pretty quick but not via the context of a public forum.

 

Kirk..

Beginner

Re: Service Profile Association Fails - "server unavailable There are not enough resources overall"

Hi Kirk-

 

I see what you a referring to and understand that if the blade is running a newer version of firmware than the infrastructure, then its not supported. In researching downgrading the blade firmware, I came across this: https://www.cisco.com/c/en/us/td/docs/unified_computing/ucs/ucs-manager/GUI-User-Guides/Firmware-Mgmt/4-0/b_UCSM_GUI_Firmware_Management_Guide_4-0/b_UCSM_GUI_Firmware_Management_Guide_4-0_chapter_010.html#concept_28CA150BA723426DAA865FD5665E9189

"he board controller firmware version of the blade server should be the same as or later than the installed software bundle version. Leaving the board controller firmware at a later version than the version that is currently running in your existing Cisco UCS environment does not violate the software matrix or TAC supportability."

I took from this to mean that I was ok to run the later version on a blade, but maybe its specific to just the board controller firmware?

 

To answer your questions:

Yes the blade fully discovered, as best as I can tell. It shows a complete inventory and no errors. I tried re-acknowledging several times, and physically removing and reseating the server.

 

I confirmed several times that my IP, MAC, UUID, etc. pools had plenty of available addresses.

 

I am not using server pools and have assigned SP's to blades manually.

 

I also tried building a new service profile just for this blade, but it fails during the association wizard before it will even apply with the "server unavailable There are not enough resources overall" error. 

 

Can you point me in the direction of the logs that I can check to help resolve the error?


Thanks for your help thus far!


Derek

Cisco Employee

Re: Service Profile Association Fails - "server unavailable There are not enough resources overall"

Board controller firmware is the one exception (as you can't downgrade it).  But your overall UCS-C or UCS-B packages need to fall within that cross-support matrix.

You'll need to open a TAC case, if you are still having issues after getting the firmware within spec.

 

Kirk...

Beginner

Re: Service Profile Association Fails - "server unavailable There are not enough resources overall"

Thanks for the help, upgrading the infrastructure to 4.02 seems to have fixed the problem. Initially I received the same error when I applied the service profile, so I disassociated it, and then applied a "dummy" service profile with minimal config, disassociated that and then reapplied the production service profile, and it took. Applying the the dummy sp seemed to do some firmware changes that the prod sp did not, so maybe that had something to do with it.

 

Appreciate the help!

CreatePlease to create content
Content for Community-Ad
July's Community Spotlight Awards