I have eight physical hosts to migrate, five are running CentOS 4.5+ and seem to be manageable. The other three are running RHEL 3.0 (no Update level).
I tried the RHEL ones first with converter version 5.1. I got an issue as pasted below from the log:
--> msg = "An error occurred during the conversion: 'GrubInstaller::InstallGrub: /usr/lib/vmware-converter/installGrub.sh failed with return code: 139, and message:
--> /mnt/p2v-src-root/sys does not exist. Not mounting
--> /usr/lib/vmware-converter/installGrub.sh: line 132: 1997 Segmentation fault chroot "$argRootPath" /$updateGrubName "$argGrubVersion" "$argBootDisk" "$argInstallPart" /$deviceMapName
--> umount: can't umount /mnt/p2v-src-root/sys: No such file or directory
--> Error running vmware-updateGrub.sh through chroot into /mnt/p2v-src-root
--> /usr/lib/vmware-converter/installGrub.sh: line 143: /mnt/p2v-src-root: Is a directory
I started debugging Grub issues, but could only find 1 vs 2 types of problems, for Ubuntu. This server says it's running GNU GRUB 0.93.
So I starting think of overall compatibility:
VMware Compatibility Guide: Guest/Host Search
The matrix seems to say only the last update level (Update 9) of RHEL 3.x is supported by ESXi 5.x (I have 5.1U1).
So now I have three courses of action:
1. Since I can't find the exact return code 139 for my problem above in Google, should I keep trying to make it work as-is? What was the big difference in RHEL 3 Update 9 that allows ESXi 5 to support it (but not Update 8 or earlier)? Was it just formal support, but it will probably still 'work'? Either way - do I have a shot if I keep debugging this Grub error?
2. Or - should I be looking to upgrade my host to RHEL 3 Update 9? (I am not sure if this is even possible, or how difficult it will be - thus my next question)
3. Or - should I be looking at doing some sort of a net dd operation, and then manually updating the devices to the ones ESXi provides virtually to try to get these to bootstrap virtually?
CAVEAT: One thing I will mention is this - I am 3,000 miles from the servers, and the staff doesn't seem able to do much beyond powering on/off. I would think of having DRACs installed, but that would also cost a lot (their hourly rate is high), and I don't want them tearing apart these legacy machines just for that. In summary, I would like to get these virtualized without powering down at all. I believe minor RHEL upgrades can be done without powering down, but I have no idea on this old version. Net dd should be fine. I just can't be relying on rebooting the physical machines and doing terminal work unless I decide I have to fly to the site to finish this.
EDIT: I would also like to add this - for my Centos machines, the one that I got to work only did so because I downgraded to Converter version 5.0.1, and converted the LVM to basic during conversion (based on other threads here). Perhaps I should be trying version 4.x of the Converter for these ancient RHEL 3.x machines?