Hey all,
I'm interested in some discussion around file system resilience.
I appreciate, that in reality, most VMFS failures we hear about involve either hardware of user error, however, those errors are still worth discussing. They are common on this forum.
With the upcoming Windows Server 2012 we have ReFS - standing for "Resilient File System" - replacing NTFS. It's built with quite a few features aimed at proactively managing the sorts of corruptions and issues we used to have on NTFS. Again - to be fair - most of my experience with such issues has gotten a lot better as underlying hardware has gotten better from the days we used to run Windows 2000.
That said, more resilience is, in my mind, always better, and I welcome the new file system. The underlying issue then is, does a virtualised machine running ReFS, benefit any, when the environment is virtualised across VMFS?
I would imagine "somewhat". If VMFS is able to see a usable VMDK enough to boot an OS, then having that OS able to repair itself of any issues is very helpful. However, issues with VMFS that either render it unable to mount under ESXi, or have it show up empty, don't help in this regard.
There have been a number of improvements in VMFS, in either the form of drivers accessing it updating between ESXi versions, or in moving from VMFS3 to VMFS5.
I think most of the community appreciates the removal of the 2TB limitation, but surely the next step is availability.
On a basic level, some sort of "chkdsk"-like command is well overdue.
The old "I corrupted my FS by opening it in Windows and now need to use fdisk to repair it" seems to be a FAQ on this forum, and it's time there was a "esxcli" command to automate this.
Again, I know if we all have quality SAN architecture this shouldn't be relevant. However on a competing side, the reality is that Hyper-V will soon be running on such a platform, and it falls to VMware to keep up.
Thoughts?