Hi, On Fri, 2006-03-31 at 19:17 -0500, Kwan Lowe wrote: > Yesterday I installed the Xen0 kernel on my primary desktop to try to test xen > migration. I built out a couple 4G logical volumes and ran a couple kickstart > installations. At one point I tried to expand my /home filesystem (it's on a > /dev/rootvg/home_lv LVM partition). It failed, and then my /home seems to have been > remounted as read-only. Long story short, on a reboot the fsck.ext3 informs that the > filesystem is larger than the disk and fsck just gives hundreds of errors. Do you have any record of exactly what went wrong --- kernel error logs, etc? It sounds as if the growth of the LV wasn't propagated to disk, which is bad news. Was /home exported to any of the guests? > I'm going > to blow away that LV and try to rebuild it tomorrow, but just curious if anyone else > has seen any disk corruption? I haven't seen *any* such problems during all the time I've been using and developing on Xen. But virtual storage does give you new ways of shooting yourself in the foot, so there may well be missing safety-nets in some of the tools surrounding Xen. We'd need to work out what went wrong first here to determine how best to guard against it. Cheers, Stephen
