I’ve have a couple of updates for my post about Dell’s SUU 5.4.0 and ESX 3.5.
First, “suu -c” will tell you what the updates you need are, but “suu -u” won’t actually apply them. Instead you get an error. While you might think “that sucks!” it’s actually fine, because…
It looks like some Dell updates for older servers cannot be applied to a system that’s been up & operational for a long time without a reboot. For me, BIOS updates would complain about not enough contiguous RAM, and RAID controller updates caused ESX to panic with a pink debugger screen of death. A reboot prior to the firmware updates appears to prevent problems.
My test environment is two Dell PowerEdge 1850s and two Dell PowerEdge 6650s, and trying to update the RAID controller firmware on three of them caused a kernel panic. The fourth had been rebooted recently and was fine, and each machine that crashed updated properly after they rebooted. Newer machines, like my farm of 2950s, have taken updates well, though. As with anything, testing first is key.
Last, Scott Hanson from Dell wrote a script to apply the SUU-recommended patches, by just getting the output of “suu -c” and running each update individually. His script works great, and since I copy the SUU DVD to an NFS export anyhow I can just add the script as my replacement for running SUU. Given what I noted above about older hosts you’ll probably want to reboot them first, though. 🙂
