we're running into a scenario similar to this with CentOS 6, zero length files after a system crash. http://oss.sgi.com/archives/xfs/2012-02/msg00607.html
I'm not exactly sure which kernel version they are running in production (its in China), but I'm trying to find out. its probably a year or so old, our manufacturing operations folks do NOT like installing random updates without very good reasons.
meanwhile, I wonder if anyone familiar with it knows if the fix that Dave Chinner discusses in that posting I link above has been backported to a recent EL6 kernel.
John R Pierce:
we're running into a scenario similar to this with CentOS 6, zero length files after a system crash. http://oss.sgi.com/archives/xfs/2012-02/msg00607.html
I'm not exactly sure which kernel version they are running in production (its in China), but I'm trying to find out. its probably a year or so old, our manufacturing operations folks do NOT like installing random updates without very good reasons.
meanwhile, I wonder if anyone familiar with it knows if the fix that Dave Chinner discusses in that posting I link above has been backported to a recent EL6 kernel.
This was fixed in EL6 kernels a while ago - see:
http://rhn.redhat.com/errata/RHSA-2012-1401.html
James Pearson
On 6/10/2013 11:43 AM, James Pearson wrote:
This was fixed in EL6 kernels a while ago - see:
ahhh, thanks. turns out they were running stock unpatched 6.2 with kernel 2.6.32-220.el6.x86_64
2.6.32-220-28.1 got this fix.
advising they update all the systems to latest.
-
James Pearson -
John R Pierce