[Solved] Newest VMWare Tools in ESXi 5 don't play well with Gentoo Linux
I get the impression that Gentoo is not widely used in the Enterprise and particularly not widely used as a Guest OS for vSphere/ESXi in the Enterprise. But, my hope is that someone out there is in our boat and can possibly help with this strange issue we are having. Or perhaps someone can help in some way anyway.
Our environment: We currently run 4 ESXi hosts on Dell M610/M1000e blade hardware with shared EqualLogic storage. All are running the latest build of ESXi 5.0.
We are running about 40-50 Hardened Gentoo Linux VM's. Mixture of 32 and 64-bit kernels. Kernel versions ranging from 2.6.29-hardened to 3.1.5-hardened.
Most VM's are hardware v8 with a spattering of v7.
The story: We used to run ESXi 4.0. We never had any issues compiling and running official VMWare Tools in Gentoo, which would gracefully handle restarts and shutdowns (amongst other features it obviously provides) from the vSphere client no problem. Since migrating our VM's to ESXi 5.0 and subsequently upgrading out of date VMWare Tools via Update Manager (22.214.171.12468 (build-446312)), we are experiencing a strange issue. In order to gracefully restart or shutdown a Gentoo guest from the vSphere client, we must issue the restart or shutdown _twice_ from the vSphere client. Nothing happens at all the first time. On the second attempt, the shutdown or restart process begins normally. You can likely understand why this would be a problem in automating tasks.
At this point I should note, that yes, we have tested open-vm-tools. They work fine and do not exhibit the above issue. However, not using VMWare distributed tools limits our ability to manage the automated updating of tools for example via VMWare Update Manager. Also, we DID test officially supported distros (like CentOS), with the same version of tools we are installing in Gentoo. They do NOT experience the issue noted above.
We also did not experience the issue on VM's migrated from ESXi 4.0 to 5.0 before we updated tools to newest version.
So, this appears be newest VMWare tools not playing well with Gentoo. Not sure which is really at fault as logs do not help (my gut feeling is tools based on the troubleshooting). The command to shutdown or restart just appears to go into a blackhole until the 2nd attempt.
It is not the end of the world, it is just annoying and inconvenient. VMWare support has attempted to help us over a few calls and WebEx sessions. They did determine that the case should be escalated to engineering, but because Gentoo is not officially supported as a guest OS, they will not do so. We are well aware that Gentoo is not officially supported, however, we have always found ways to make it work well for us until now. Our use of hardened Gentoo stems back to pre-VM days as a corporate policy and will likely not change anytime soon, so we would prefer to attempt to solve this issue as using VMWare distributed tools is the best case scenario for us. It seems like something changed in the new tools that would be easy enough to resolve if we could only get the ear of someone kind in Engineering who can make an exception for the "unsupported" nature of our guest OS.
Thanks for any help. Let me know if I can provide more information!