cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1876
Views
1
Helpful
6
Replies

UCS Director - Reserve IP in Static Pool

joel.cason
Level 1
Level 1

I know the best idea here is likely to reserve an IP block per subnet that UCSD doesn't control, but is it possible to reserve an IP somehow such that a static pool won't allocate it, other than excluding it from the list?

E.g. if I create two VMs with UCSD, but need to reserve a VIP in the same pool, can I mark the IP as used even though UCSD doesn't create it?

Or, is it possible to modify an existing, in-use static pool without affecting existing reservations?  So I can modify the IP range values to exclude an IP that is in use?

1 Accepted Solution

Accepted Solutions

I would advise making an IP address reservation compound workflow.  You could have it cycle through addresses until it finds one that is not in use.  IIRC, “Static IP Pools” will skip IPs that UCSD has seen in use from inventories of your vCenter.  E.g. someone builds a VM with 192.168.5.55, and your pool is 5.0/24, 55 ‘should’ be skipped.  However, if baremetal uses that, UCSD would have no way to know that was the case.    So just to be safe, I’d have a workflow do the work of pinging and if it is in a subnet where ‘ping’ is blocked, you might need to have the gateway of the subnet do the ping and check for an ARP entry to confirm it is free.  So what you ask for is doable, just that it’ll take some validation and you’ll have to put in a little error checking and not just rely on the single task.

View solution in original post

6 Replies 6

Orf Gelbrich
Cisco Employee
Cisco Employee

There is a task to pull an IP from a POOL.

OK I see this task (Get IP Address from Pool) and it tests out good, but this would be for pulling an additional random address out of the pool correct?

So I could provide an additional IP if someone asks for it, but if someone assigns an IP and then lets me know what it is, is there a built in way to mark it as reserved? 

To the best of my knowledge once you pull an IP with this task it will not pull the same IP again.

I would advise making an IP address reservation compound workflow.  You could have it cycle through addresses until it finds one that is not in use.  IIRC, “Static IP Pools” will skip IPs that UCSD has seen in use from inventories of your vCenter.  E.g. someone builds a VM with 192.168.5.55, and your pool is 5.0/24, 55 ‘should’ be skipped.  However, if baremetal uses that, UCSD would have no way to know that was the case.    So just to be safe, I’d have a workflow do the work of pinging and if it is in a subnet where ‘ping’ is blocked, you might need to have the gateway of the subnet do the ping and check for an ARP entry to confirm it is free.  So what you ask for is doable, just that it’ll take some validation and you’ll have to put in a little error checking and not just rely on the single task.

I forgot to mention that as you pull IPs to test, if they are in use, I’d leave those reservations in so the next time you don’t use them and Director won’t offer them up.

I used this recently in a workflow (https://communities.cisco.com/docs/DOC-60468) to test for port 443. When it is up I have the workflow move on to the next part. First I test if ESXi is up and the I test if my vCenter installed.

Review Cisco Networking for a $25 gift card

Cisco UCS X-Series Energy Efficiency Offer