iSCSI boot vNICs pingable during OS install, but not during boot

[derek.small]

I've configured boot from SAN on UCS with both iSCSI and FC a number of times, but either I'm getting rusty or something has changed.  UCS cluster is running 4.1.  I have two older M4 blades that I'm trying to boot from a Nimble iSCSI SAN.  After creating my service profiles and assigning them to the blades, then booting the blades.  I see the iSCSI bios get installed during boot, and it sees the Nimble as my target.  I'm installing VMware ESX7.0U3, and I've tried both the VMware vanilla ISO and the Cisco specific one.  I can install the OS, and I see connections on my Nimble while the installation is taking place, and the LUN I created gets something transfered to it during the install.  I also created a temporary SVI on my switch to the iSCSI Vlan and I can ping the iSCSI vNICs on my blade while the OS is installing.  The installation process has to get passed the screen where it's loading modules (to the screen that is half yellow and half black if you know the ESX installer).  The install seems to progress normally, but after eveything is done once I reboot the blade, it won't boot from SAN.

While trying to boot, I cannot ping the iSCSI vNICs on my blade/host.  I also find it troubling that I cannot ping the iSCSI vNICs until I get pretty into the installer.  For these to work as boot from SAN, they should be pingable as soon as the server is powered up.  I cannot figure out where the problem is though.  The iSCSI vNICs are pulling IP addresses from the IP pool I configured for them, they just don't seem to be doing it early enough in the boot process.  I'm also confused how the iSCSI BIOS seems to be loading during boot up and connecting to the Nimble, but I cannot ping my iSCSI vNICs during that process.

Anyone have any ideas where to look?

[Cisco_Virtual_Engineer]

Based on the information you provided, I can suggest a few areas to investigate and some possible solutions:

1. Double-check the UCS iSCSI Boot Configuration:
Ensure that the UCS iSCSI boot configuration is correctly set up. This includes verifying the iSCSI vNICs, boot targets, and LUN settings in the service profile. Also, ensure that you have the correct boot order, with iSCSI boot as the primary option.

2. Verify the iSCSI VLAN Configuration:
Make sure that the iSCSI VLAN is properly configured on both the UCS Fabric Interconnects and the upstream switches. Ensure that the VLAN is allowed on all relevant interfaces and that proper tagging is in place.

3. Confirm the iSCSI Initiator and Target Settings:
Check the iSCSI initiator settings on the UCS blades and ensure they match the target settings on the Nimble SAN. This includes confirming the iSCSI Qualified Name (IQN) and authentication settings if required.

4. Update Firmware and Drivers:
Ensure you are using the latest firmware and drivers for the UCS blades, Fabric Interconnects, and Nimble SAN. Compatibility issues between different firmware and driver versions can cause unexpected behavior.

5. Check the Boot LUN Configuration on the Nimble SAN:
Verify that the Boot LUN is correctly configured on the Nimble SAN, and ensure that it is mapped to the correct iSCSI initiator IQN. Make sure the LUN is set to be bootable.

6. Review ESXi Installation Logs:
After the installation of ESXi, review the installation logs (e.g., /var/log/esx_install.log) for any errors or issues related to iSCSI boot or connectivity.

7. Verify BIOS Settings on UCS Blades:
Ensure that the UCS blade BIOS settings are configured to support iSCSI boot. This may include enabling the iSCSI boot option in the BIOS and configuring the correct boot order.

If you've already checked these areas and are still experiencing issues, I recommend opening a support case with Cisco TAC for further assistance. They can help you with in-depth troubleshooting and provide insights specific to your environment.