VMware Cloud Community
rock_larson
Contributor
Contributor

Re-ip'ing the ESX Cluster setup + VMs

Hi,

We are moving our VLAN completely from one ip series to another ip series for which we need to re-ip all our ESX servers and VMs hosted on them. We have ESX 3.0.1 (without patches) with both DRS and HA cluster enabled. I have below plan in mind and hoping that this is work, could someone please suggest if any corrections are required in below plan if they have already

done it??

*Shutdown & Power off all the VMs

*Put the ESX servers into Maintenance Mode

*Reassign new IP range to ESX servers - http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&externalId=4309499 *Reboot the ESX servers without storage and change the storage target IPS, reboot

*Verify that both ESX server now see storage

*Bring the servers out of maintenance mode and start powering on the VMs, without network

*Re-assign the new IP to the VMs

*Verify that ESX Cluster is fine

Thanks in Advance and sorry for asking too many questions...

0 Kudos
8 Replies
rock_larson
Contributor
Contributor

no suggestions or comments from anyone?. May i consider my plan as fine ?

0 Kudos
whynotq
Commander
Commander

I saw a problem for a customer of mine doing this, I suggest you disable HA before making these changes. They went through your plan almost identically and the only problem cam when they tried to re-enable HA at the end of it. It was a problem with DNS which meant that although the Hosts could resolve and VMKping worked HA just kept failing. I have since tried it and disabling HA firt seems to work.

0 Kudos
rogilvie
Contributor
Contributor

\- as previously stated, disable HA until your finished - just confuses things

\- make sure all DNS records have updated

\- might want to update your hosts file on the ESX boxes

\- i know its obvious, but remember to update your VM network with the new IP!

here is some good doco if you run into problems changing the IPs of the ESX boxes, just delves further in the console networking commands, module 4 networking: http://www.rtfm-ed.co.uk/?p=261

Don't know how much time you have, once you have migrated the vm's successfully over to the new IPs you might want to think about applying some of the patches, even if you just apply the critical patches your environment requires.

0 Kudos
rock_larson
Contributor
Contributor

thanks for the inputs....can anyone help me to know how to disable the h/w iscsi temporarily...i know that s/w iscsi can be enabled/disabled using the commands 'esxcfg-swiscsi -e' / 'esxcfg-swiscsi' -d..i couldn't find an equivalent option for h/w iscsi

0 Kudos
Erik_Zandboer
Expert
Expert

It is considered best practice to add the ESX hostnames to the /etc/hosts file. This is exactly why whynotq sawthe problem with HA; you do not want HA to activate just because there is an error in DNS (or DNS is not avail).

Visit my blog at http://www.vmdamentals.com
0 Kudos
Erik_Zandboer
Expert
Expert

Why would you want to disable iSCSI?

Visit my blog at http://www.vmdamentals.com
0 Kudos
rock_larson
Contributor
Contributor

Hi Erik,

I wanted to disable iscsi because even my storage servers IPs are being migrated to the same VLAN as my ESX servers. So the iscsi targets need to be changed in the dynamic discovery. I thought of disabling storage and then change the ip of ESX SC1, SC2 and VMKernal and then change storage IP and enabled iscsi. Isn't it a right way of doing to avoid any kind of server hangs while my ESX searches for the VMFS luns where the target is not available at all??

Thanks

0 Kudos
Erik_Zandboer
Expert
Expert

Aha, now I understand. Yes indeed, if you are migrating iSCSI as well that would be problematic. Still, I guess you planned downtime for the VMs situated on the iSCSI storage, so it would not really matter if iSCSI went "lost" on an esx host? You could simply remove the storage, change all IP adresses, rescan, add storage.

Visit my blog at http://www.vmdamentals.com
0 Kudos