Bugzilla – Bug 257
XFS umount leave a binary file in place off /etc/fstab and /etc/mtab, which cause LInux to be unable to boot.
Last modified: 2003-06-30 14:09:31 CDT
I think it is a problem with unmount since /etc/fstab is replace by a binary files instead off a test config files with meaning. I fix this problem by rebooting from cd rom and replace fstab with a right fstab.
Created attachment 82 [details] repair seasion afte fixing /etc/fstab on xfs root partino
We're going to need a lot more information than you have provided, to get anywhere with this bug. For starters, what kernel are you running, and where did the xfs code come from. Is this reproducible? If so, how? If not, what were you doing before you saw the problem? What did this "binary file" look like? With the bug as you have filed it, for all I know you copied /proc/kcore to /etc/fstab :)
Unfortunately the attachment does not tell us very much. This part: Phase 4 - check for duplicate blocks... - setting up duplicate extent list... - clear lost+found (if it exists) ... - clearing existing "lost+found" inode - marking entry "lost+found" to be deleted tells us that there was an existing lost+found/ directory, and that it was cleared. Anything that -was- there is later re-found in later phases. So, it is very likely that the repair output has nothing to do with the problem you saw.
There was a lost+found dir before I ran xfs_repair.
I think this bug is somehowe relative to bug 258, since I alsao have highio and highmemory enabled.
We cannot do anything with this bug until you provide some basic information as I have requested above. If you cannot provide any more information about what you were doing, or what your "fstab" looked like, or even what kernel you were running, I will have to close this bug as WORKSFORME.
I am unable to reproduce since I have turn off highmemio and highmemory. /etc/fstab contain just meanlinless binary. *** This bug has been marked as a duplicate of 258 ***
> I am unable to reproduce since I have turn off highmemio and highmemory. And you -can- reproduce it when they are turned on? If so, how? bug 258 is related to the XFS_IOC_RESERVSP ioctl, are you using that call? It also relates to a system lockup - did your system lock up before you saw your problems? Un-duping and closing this bug until we get some useful information.