ciscoucs / kubam Goto Github PK

NTP is not properly starting preventing the ansible play book from completeing. I had to ssh to each node and manually restart ntp and manually update it with the proper server.

when creating boot image the template may not have an SSH key so should make it not required.

Hi guys,
Just upgraded UCSM and got this error:
Unsupported UCS firmware version: 4.0(1a). Please update to at least 3.0

Looks like CentOS 7.3 is hardcoded in the environment file. Needs to be updated to be dynamic.

Requested by Cisco Customer for generic bare metal UCS provisioning.

If the UCS only has one fabric interconnect configured make sure KUBAM is smart enough and only configures one network instead of 2 different VLANs or it doesn't work.

My iso for 6.7 gets created in the /home/%user%/kubam folder but when I mount it to the iso either via vmedia or just through a standard KVM connection it's not bootable. Am I supposed to do something special to make that iso bootable. If I mount it in the OS it looks the same as the normal esx 6.7 iso other than I see the new KS.CFG file I expected to see.

I also have the same issue with esx 7

If NTP is way way way off then the kubernetes and docker repositories will complain about bad certificate. This has to do with the nodes being out of sync on time rather than bad repositories. To fix this, put the NTP configuration in the common Ansible role at the beginning before adding these repositories.

Hi - the docs site is a bit old, still mentions March updating... https://ciscoucs.github.io/site/

Make sure that each host has the same OS if they are in the same server group

you can name the network somethign with spaces or dashes but should force it to be a name that is url safe.

Is it possible to modify the kubeucs.py script to automatically run the resetDisks.py script, if the RAID LUN is orphaned?

The front end calls /isos/boot API which in turn calls the API function mkboot_iso() which calls: IsoMaker.mkboot_iso and Builder.deploy_server_images.

mkboot_iso seems to work as it extracts the ESXi image, but no images are created for the nodes. This is probably in Builder.deploy_server_images.

power stat command always returns off. Need to investigate this. Tested with UCS Manager but UCS Central may not have been tested.

Running stage2 python script gets the error:

# ./kubeucs.py -o micdoher1425 admin nbv12345 172.28.225.163
Checking if org micdoher1425 exists
Org micdoher1425 does not exist
Creating Organization: micdoher1425
[1] VLAN default
[2] VLAN hx-inband-mgmt
[3] VLAN hx-vmotion
[4] VLAN hx-data
[5] VLAN Docker-Data
[6] VLAN Docker-storage
[7] VLAN hx-storage-data
[8] VLAN vm-network
--------------------------------------------------------------------------------
Please Select a VLAN for the Kubernetes Server to use: 1
Creating Kubernetes MAC Pools
org: org-root/org-micdoher1425
Creating Kubernetes VNIC Templates
Creating Kubernetes LAN connectivity policy
Creating Kube Boot Policy
Creating Kube Bios policy
Adding Disk Group Config
Creating Kube UUID Pools
Creating Kubernetes Compute Pool
Adding Virtual Media Policy
What is the URL for the Boot ISO image?
(E.g.: http://192.168.2.2/kubam/centos7.2-boot.iso) : http://172.28.225.60/kubam/centos7.3-boot.iso
You entered: http://172.28.225.60/kubam/centos7.3-boot.iso
Is this correct? [y/N]: y
Listing Available UCS Servers
[1]: Blade 1/slot-1 type UCSB-B200-M4
[2]: Blade 1/slot-3 type UCSB-B200-M3
[3]: Blade 1/slot-5 type UCSB-B200-M3
[4]: Blade 1/slot-6 type UCSB-B200-M3
[5]: Blade 1/slot-7 type UCSB-B200-M3
[6]: Blade 1/slot-8 type UCSB-B200-M3
[7]: Rack rack-unit-1 type HX240C-M4SX
[8]: Rack rack-unit-2 type HX240C-M4SX
[9]: Rack rack-unit-3 type HX240C-M4SX
[10]: Rack rack-unit-4 type HX240C-M4SX
[11]: Rack rack-unit-7 type UCSC-C220-M3S
[12]: Rack rack-unit-8 type UCSC-C220-M3S
[13]: Rack rack-unit-9 type UCSC-C220-M3S
[14]: Rack rack-unit-11 type UCSC-C240-M4SX
[15]: Rack rack-unit-12 type UCSC-C240-M4SX
[16]: Rack rack-unit-6 type UCSC-C220-M3S
[17]: Rack rack-unit-10 type UCSC-C240-M4SX
[18]: Rack rack-unit-5 type UCSC-C220-M3S
(E.g.: 2,4,8): 16,11,12
Install Kubernetes on the following servers:
            Server rack-unit-6 type UCSC-C220-M3S
            Server rack-unit-7 type UCSC-C220-M3S
            Server rack-unit-8 type UCSC-C220-M3S
Is this correct? [N/y]: y
Adding servers to Kubernetes Pool
Traceback (most recent call last):
  File "./kubeucs.py", line 59, in <module>
    main()
  File "./kubeucs.py", line 42, in main
    UCSServer.createKubeServers(handle, org)
  File "/root/Code/KUBaM/stage2/server/UCSServer.py", line 521, in createKubeServers
    addServersToKubePool(handle, servers, org)
  File "/root/Code/KUBaM/stage2/server/UCSServer.py", line 226, in addServersToKubePool
    reset_disks(handle, s)
  File "/root/Code/KUBaM/stage2/server/UCSServer.py", line 78, in reset_disks
    for d in disks:
TypeError: 'NoneType' object is not iterable

i cannot many any ISO using KUBAM web interface. i have uploaded the ISO(wget centos.iso) inside kubam docker images but i cannot see anything to map iso and this is show stopper to me.

Can anyone help.

BR
Rushi

API to set KUBAM IP may be broken.

Hello,

First of all, I would like to thank you for the work you've done with KuBAM, it's a great tool !
I'm trying to make it work with ESXi (6.7 U1), and I have an issue with the vMedia policy.
As seen in the documentation, I configured a CDD one, with the service-profile-name as image parameter. Unfortunately it does not seem to work (iso file is $service-profile-name .iso). When I set it manually (complete path), it seems to work.

Could you please help me resolve this ?

Thanks,

Tristan

Need to create a vMedia Policy for ESXi that does the boot a different way. Current way tries to do it like CentOS and fails.